agentic-awesome-skills
Flagged as possibly malicious Reviewed: expected behaviour
Bawwab's static rules matched a pattern it treats as critical. Read the review notes: a reviewed false alarm is marked as such.
Why it was flagged
The static rules marked this “Flagged as possibly malicious” because of:
exfil.combo: Reads credentials and makes network calls (critical, 2× · first at skills/liuguang-banlan-ui/tests/test_review_boundaries.py:76)
- Kind
- Plugin
- Source
- github
- Repository
- sickn33/agentic-awesome-skills / plugins/agentic-awesome-skills-claude
- Commit
4a52ab5ef117bdce2c6ba8d87dd64df0c840dfea- Listed in
- sickn33/agentic-awesome-skills
- Popularity
- 50,025 GitHub stars (2026-11-20)
- Link
- https://github.com/sickn33/agentic-awesome-skills/tree/4a52ab5ef117bdce2c6ba8d87dd64df0c840dfea/plugins/agentic-awesome-skills-claude
- Scanned
- 2026-11-20T23:08:50+00:00
- Scan time
- 188.1 s
- Scanned as
- plugin
Review note: Large skill collection; the flagged lines are scanner code, documentation, help text and security-testing examples, with no code found that reads secrets and sends them elsewhere. One skill to know about: movie-coupon fetches a third-party advertising feed (agskills.moontai.top) into your agent; the skill itself labels that content untrusted.
Findings (69)
files.truncated: Scan limit reached med
file limit of 5000 reached; remaining files not scanned
What this means: Some files, or the tail of a very large file, were not scanned, so malicious code could hide in them.
The rule's general advice: Review the unscanned files manually or raise the limits. (see the review note for this case)
Expected for what this tool does
This collection is larger than Bawwab's 5,000-file scan limit, so part of it was not scanned. Our hand-check covered only the files with findings; the unscanned files are not verified.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/007/scripts/config.py:312
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
".json", ".env", ".env.example",
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
A list of file extensions (including .env) that the 007 audit scanner is willing to scan; nothing is read or sent.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/007/scripts/full_audit.py:247
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
"Secrets Manager, HashiCorp Vault), or a .env file excluded from "
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
Remediation advice string telling users to keep secrets in a .env file excluded from version control.
Reviewed by Bawwab maintainers, 2026-11-20
remote-code.pipe-shell: Pipes a download into a shell high
skills/007/scripts/scanners/dependency_scanner.py:807
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesOWASP Agentic ASI05 Unexpected Code Execution (RCE)MITRE ATLAS AML.T0050 Command and Scripting InterpreterCWE CWE-494 Download of Code Without Integrity Check
description="Pipe-to-shell pattern detected (curl|bash). Remote code execution risk", # security-allowlist: curl-pipe-bash
What this means: Executes whatever the remote server returns, unreviewed.
The rule's general advice: Download, verify a checksum, then run. (see the review note for this case)
Known false alarm
The scanner reports curl|bash patterns it finds in scanned projects; this line is the description text of that report, not a pipe-to-shell command.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/007/scripts/scanners/dependency_scanner.py:166
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
r"""^\s*(?:COPY|ADD)\s+.*?(?:\.env|\.key|\.pem|\.p12|\.pfx|id_rsa|id_ed25519|\.secret)""",
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
Regex used to flag Dockerfile COPY/ADD of sensitive files in scanned projects.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/007/scripts/scanners/secrets_scanner.py:5
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
file types (.env, config files, shell scripts, Docker, CI/CD).
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
Module docstring of a secrets scanner describing the file types it checks.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/ai-studio-image/scripts/config.py:32
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
2. .env GEMINI_API_KEY
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Expected for what this tool does
Image-generation skill looks up its own GEMINI_API_KEY from the environment or a local .env as documented in the docstring; the key is used for the Gemini API.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/ai-studio-image/scripts/generate.py:286
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
print(" 2. Arquivo .env em: C:\\Users\\renat\\skills\\ai-studio-image\\")
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
A help message printed when no API key is set, naming where to put the key.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/apify-audience-analysis/reference/scripts/run_actor.js:75
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
The match is a usage line in the help text (node --env-file=.env ...). The script reads APIFY_TOKEN from the environment and sends it only to api.apify.com.
Reviewed by Bawwab maintainers, 2026-11-20
exfil.combo: Reads credentials and makes network calls med
skills/apify-audience-analysis/reference/scripts/run_actor.js:75
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0037 Data from Local SystemMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0025 Exfiltration via Cyber Means
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Credential access together with outbound requests in one file is the shape of data theft.
The rule's general advice: Do not install unless the flow is understood and benign. (see the review note for this case)
Known false alarm
Same help-text line. The script's only network calls go to api.apify.com with the user's own APIFY_TOKEN and redirects disabled; no other host is contacted.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/apify-brand-reputation-monitoring/reference/scripts/run_actor.js:75
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
The match is a usage line in the help text (node --env-file=.env ...). The script reads APIFY_TOKEN from the environment and sends it only to api.apify.com.
Reviewed by Bawwab maintainers, 2026-11-20
exfil.combo: Reads credentials and makes network calls med
skills/apify-brand-reputation-monitoring/reference/scripts/run_actor.js:75
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0037 Data from Local SystemMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0025 Exfiltration via Cyber Means
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Credential access together with outbound requests in one file is the shape of data theft.
The rule's general advice: Do not install unless the flow is understood and benign. (see the review note for this case)
Known false alarm
Same help-text line. The script's only network calls go to api.apify.com with the user's own APIFY_TOKEN and redirects disabled; no other host is contacted.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/apify-competitor-intelligence/reference/scripts/run_actor.js:70
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
The match is a usage line in the help text (node --env-file=.env ...). The script reads APIFY_TOKEN from the environment and sends it only to api.apify.com.
Reviewed by Bawwab maintainers, 2026-11-20
exfil.combo: Reads credentials and makes network calls med
skills/apify-competitor-intelligence/reference/scripts/run_actor.js:70
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0037 Data from Local SystemMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0025 Exfiltration via Cyber Means
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Credential access together with outbound requests in one file is the shape of data theft.
The rule's general advice: Do not install unless the flow is understood and benign. (see the review note for this case)
Known false alarm
Same help-text line. The script's only network calls go to api.apify.com with the user's own APIFY_TOKEN and redirects disabled; no other host is contacted.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/apify-content-analytics/reference/scripts/run_actor.js:75
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
The match is a usage line in the help text (node --env-file=.env ...). The script reads APIFY_TOKEN from the environment and sends it only to api.apify.com.
Reviewed by Bawwab maintainers, 2026-11-20
exfil.combo: Reads credentials and makes network calls med
skills/apify-content-analytics/reference/scripts/run_actor.js:75
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0037 Data from Local SystemMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0025 Exfiltration via Cyber Means
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Credential access together with outbound requests in one file is the shape of data theft.
The rule's general advice: Do not install unless the flow is understood and benign. (see the review note for this case)
Known false alarm
Same help-text line. The script's only network calls go to api.apify.com with the user's own APIFY_TOKEN and redirects disabled; no other host is contacted.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/apify-ecommerce/reference/scripts/run_actor.js:70
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
The match is a usage line in the help text (node --env-file=.env ...). The script reads APIFY_TOKEN from the environment and sends it only to api.apify.com.
Reviewed by Bawwab maintainers, 2026-11-20
exfil.combo: Reads credentials and makes network calls med
skills/apify-ecommerce/reference/scripts/run_actor.js:70
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0037 Data from Local SystemMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0025 Exfiltration via Cyber Means
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Credential access together with outbound requests in one file is the shape of data theft.
The rule's general advice: Do not install unless the flow is understood and benign. (see the review note for this case)
Known false alarm
Same help-text line. The script's only network calls go to api.apify.com with the user's own APIFY_TOKEN and redirects disabled; no other host is contacted.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/apify-influencer-discovery/reference/scripts/run_actor.js:75
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
The match is a usage line in the help text (node --env-file=.env ...). The script reads APIFY_TOKEN from the environment and sends it only to api.apify.com.
Reviewed by Bawwab maintainers, 2026-11-20
exfil.combo: Reads credentials and makes network calls med
skills/apify-influencer-discovery/reference/scripts/run_actor.js:75
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0037 Data from Local SystemMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0025 Exfiltration via Cyber Means
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Credential access together with outbound requests in one file is the shape of data theft.
The rule's general advice: Do not install unless the flow is understood and benign. (see the review note for this case)
Known false alarm
Same help-text line. The script's only network calls go to api.apify.com with the user's own APIFY_TOKEN and redirects disabled; no other host is contacted.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/apify-lead-generation/reference/scripts/run_actor.js:75
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
The match is a usage line in the help text (node --env-file=.env ...). The script reads APIFY_TOKEN from the environment and sends it only to api.apify.com.
Reviewed by Bawwab maintainers, 2026-11-20
exfil.combo: Reads credentials and makes network calls med
skills/apify-lead-generation/reference/scripts/run_actor.js:75
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0037 Data from Local SystemMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0025 Exfiltration via Cyber Means
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Credential access together with outbound requests in one file is the shape of data theft.
The rule's general advice: Do not install unless the flow is understood and benign. (see the review note for this case)
Known false alarm
Same help-text line. The script's only network calls go to api.apify.com with the user's own APIFY_TOKEN and redirects disabled; no other host is contacted.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/apify-market-research/reference/scripts/run_actor.js:75
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
The match is a usage line in the help text (node --env-file=.env ...). The script reads APIFY_TOKEN from the environment and sends it only to api.apify.com.
Reviewed by Bawwab maintainers, 2026-11-20
exfil.combo: Reads credentials and makes network calls med
skills/apify-market-research/reference/scripts/run_actor.js:75
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0037 Data from Local SystemMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0025 Exfiltration via Cyber Means
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Credential access together with outbound requests in one file is the shape of data theft.
The rule's general advice: Do not install unless the flow is understood and benign. (see the review note for this case)
Known false alarm
Same help-text line. The script's only network calls go to api.apify.com with the user's own APIFY_TOKEN and redirects disabled; no other host is contacted.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/apify-trend-analysis/reference/scripts/run_actor.js:75
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
The match is a usage line in the help text (node --env-file=.env ...). The script reads APIFY_TOKEN from the environment and sends it only to api.apify.com.
Reviewed by Bawwab maintainers, 2026-11-20
exfil.combo: Reads credentials and makes network calls med
skills/apify-trend-analysis/reference/scripts/run_actor.js:75
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0037 Data from Local SystemMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0025 Exfiltration via Cyber Means
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Credential access together with outbound requests in one file is the shape of data theft.
The rule's general advice: Do not install unless the flow is understood and benign. (see the review note for this case)
Known false alarm
Same help-text line. The script's only network calls go to api.apify.com with the user's own APIFY_TOKEN and redirects disabled; no other host is contacted.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/apify-ultimate-scraper/reference/scripts/run_actor.js:70
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
The match is a usage line in the help text (node --env-file=.env ...). The script reads APIFY_TOKEN from the environment and sends it only to api.apify.com.
Reviewed by Bawwab maintainers, 2026-11-20
exfil.combo: Reads credentials and makes network calls med
skills/apify-ultimate-scraper/reference/scripts/run_actor.js:70
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0037 Data from Local SystemMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0025 Exfiltration via Cyber Means
node --env-file=.env scripts/run_actor.js --actor ACTOR_ID --input '{}'
What this means: Credential access together with outbound requests in one file is the shape of data theft.
The rule's general advice: Do not install unless the flow is understood and benign. (see the review note for this case)
Known false alarm
Same help-text line. The script's only network calls go to api.apify.com with the user's own APIFY_TOKEN and redirects disabled; no other host is contacted.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/boring-marketing-creative/scripts/verify_models.sh:22
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
if [ -f ./.env ]; then
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Expected for what this tool does
Script reads REPLICATE_API_TOKEN from ./.env when not set in the environment, to call the Replicate API for model checks.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/boring-marketing-start-here/scripts/brand_context.py:326
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
print(f"## Keys\nbrand .env present: {'yes' if (brand_dir / '.env').is_file() else 'no'} (per-client keys override the workspace .env; values are never shown)")
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
Only checks whether a brand .env file exists and prints yes/no; the text says values are never shown.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/boring-marketing-start-here/scripts/capability.py:9
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
then the resolved brand directory's .env file. Later sources take precedence.
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
Docstring describing env resolution order; it states key values are never printed.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/deploy-to-vercel/resources/deploy-codex.sh:209
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
--exclude='.env' \
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
The deploy script excludes .env files from the staged tarball, which keeps them out of the upload.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/deploy-to-vercel/resources/deploy.sh:213
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
--exclude='.env' \
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
The deploy script excludes .env files from the staged tarball, which keeps them out of the upload.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/last30days/scripts/lib/env.py:8
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
CONFIG_FILE = CONFIG_DIR / ".env"
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Expected for what this tool does
The last30days skill loads its own API keys from ~/.config/last30days/.env to call the search APIs it advertises (OpenAI, xAI).
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/last30days/scripts/lib/render.py:70
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
lines.append("- Edit `~/.config/last30days/.env` to add keys")
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
User-facing message telling the user to edit ~/.config/last30days/.env to add keys.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/last30days/scripts/lib/ui.py:95
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
{Colors.DIM}Setup:{Colors.RESET} Edit {Colors.BOLD}~/.config/last30days/.env{Colors.RESET}
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
User-facing setup banner telling the user to edit ~/.config/last30days/.env.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/launch-check/scripts/launch_check.py:15
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
HTTPS redirect, cookie flags, mixed content. With --repo: committed .env files and secrets in the repo.
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
Docstring saying the launch-check tool can look for committed .env files in a repo; it is a description of a check.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/liuguang-banlan-ui/tests/test_review_boundaries.py:76
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
(root/'.env').write_text('secret')
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
A unit test writes a dummy .env containing the word 'secret' into a temp directory to check that a local preview server refuses to serve it; nothing is read or sent.
Reviewed by Bawwab maintainers, 2026-11-20
exfil.combo: Reads credentials and makes network calls critical
skills/liuguang-banlan-ui/tests/test_review_boundaries.py:76
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0037 Data from Local SystemMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0025 Exfiltration via Cyber Means
(root/'.env').write_text('secret')
What this means: Credential access together with outbound requests in one file is the shape of data theft.
The rule's general advice: Do not install unless the flow is understood and benign. (see the review note for this case)
Known false alarm
A unit test writes a dummy .env containing the word 'secret' into a temp directory to check that a local preview server refuses to serve it; nothing is read or sent.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/macos-spm-app-packaging/assets/templates/setup_dev_signing.sh:52
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
security import "$P12_PATH" -k ~/Library/Keychains/login.keychain-db \
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Expected for what this tool does
Dev-signing helper imports a locally generated self-signed code-signing certificate into the user's login keychain, as the packaging skill describes.
Reviewed by Bawwab maintainers, 2026-11-20
install.custom-index: Installs packages from a non-PyPI index med
skills/remote-gpu-trainer/scripts/setup-china-mirrors.sh:32
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0010.001 Supply Chain Compromise: SoftwareCWE CWE-829 Inclusion of Functionality from Untrusted Control Sphere
installs Python packages from https://pypi.tuna.tsinghua.edu.cn/simple
What this means: A script runs pip install (or uv pip install) with --index-url, --extra-index-url or -i pointing at another server, so the code it installs comes from a source Bawwab never scanned.
The rule's general advice: Install from PyPI with pinned versions, or review the index and packages yourself. (see the review note for this case)
Expected for what this tool does
Script sets PIP_INDEX_URL to the Tsinghua PyPI mirror for use in China, as its name and comments say. It redirects package installs to a third-party mirror, which users should know.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/telegram/assets/boilerplate/nodejs/src/index.ts:10
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
console.error('ERROR: Set TELEGRAM_BOT_TOKEN in .env file');
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
Boilerplate bot reads TELEGRAM_BOT_TOKEN from the environment and prints an error if it is missing.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/telegram/scripts/setup_project.py:51
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
print(" 3. Copy .env.example to .env and add your bot token")
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
A next-steps message printed by the project scaffolder.
Reviewed by Bawwab maintainers, 2026-11-20
credential-access: Touches credentials or secrets high
skills/whatsapp-cloud-api/scripts/setup_project.py:81
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
env_example = os.path.join(target_path, ".env.example")
What this means: Reads files or environment holding keys, tokens or sessions.
The rule's general advice: Confirm the tool needs this; otherwise do not install. (see the review note for this case)
Known false alarm
Scaffolder copies the template's .env.example to .env in the new project; it reads no real secrets and makes no network call.
Reviewed by Bawwab maintainers, 2026-11-20
exfil.host: References an exfiltration-style host high
skills/whatsapp-cloud-api/scripts/setup_project.py:124
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0025 Exfiltration via Cyber Means
print(" ngrok http 3000")
What this means: Paste, webhook and tunnel services are commonly used to receive stolen data.
The rule's general advice: Remove or justify the destination. (see the review note for this case)
Known false alarm
The line is a printed suggestion that the user run ngrok for local webhook testing; the script does not run it.
Reviewed by Bawwab maintainers, 2026-11-20
exfil.combo: Reads credentials and makes network calls critical
skills/whatsapp-cloud-api/scripts/setup_project.py:81
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0037 Data from Local SystemMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0025 Exfiltration via Cyber Means
env_example = os.path.join(target_path, ".env.example")
What this means: Credential access together with outbound requests in one file is the shape of data theft.
The rule's general advice: Do not install unless the flow is understood and benign. (see the review note for this case)
Known false alarm
Same .env.example copy inside the new project directory; no secret is read or sent anywhere.
Reviewed by Bawwab maintainers, 2026-11-20
injection.covert-action: Tells the AI to read or send sensitive data high
skills/2slides-ppt-generator/SKILL.md:37
OWASP LLM LLM01 Prompt InjectionOWASP LLM LLM02 Sensitive Information DisclosureOWASP Agentic ASI01 Agent Goal HijackOWASP Agentic ASI02 Tool Misuse and ExploitationMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0110.000 AI Agent Tool Poisoning: Definition and InstructionsMITRE ATLAS AML.T0086 Exfiltration via AI Agent Tool InvocationCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
send it as a bearer/`apikey` value to `https:
What this means: Instructions to read keys or secrets, include them in replies or send them somewhere, or to do a hidden step first, are how injected text steals data.
The rule's general advice: Do not install unless you can explain why this text needs those files. (see the review note for this case)
Expected for what this tool does
A Security Notes section saying the API key comes from an env var, must not be echoed back, and is sent only to 2slides.com over HTTPS. It limits the behaviour rather than directing it.
Reviewed by Bawwab maintainers, 2026-11-20
injection.override: Tries to override the AI's instructions high
skills/agent-evals/SKILL.md:85
OWASP LLM LLM01 Prompt InjectionOWASP LLM LLM07 System Prompt LeakageOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0110.000 AI Agent Tool Poisoning: Definition and InstructionsMITRE ATLAS AML.T0069.002 Discover LLM System Information: System PromptCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
print system prompt
What this means: Text read by the AI that says to ignore earlier instructions, reveal the system prompt or hide things from you is a classic prompt-injection attempt.
The rule's general advice: Do not install; the text is trying to steer the assistant against you. (see the review note for this case)
Known false alarm
A test-case fixture in an evals example: a prompt that tests whether a model refuses to reveal its system prompt. It is not directed at the reader.
Reviewed by Bawwab maintainers, 2026-11-20
injection.override: Tries to override the AI's instructions high
skills/ai-red-teaming/SKILL.md:79
OWASP LLM LLM01 Prompt InjectionOWASP LLM LLM07 System Prompt LeakageOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0110.000 AI Agent Tool Poisoning: Definition and InstructionsMITRE ATLAS AML.T0069.002 Discover LLM System Information: System PromptCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
Ignore all previous instructions
What this means: Text read by the AI that says to ignore earlier instructions, reveal the system prompt or hide things from you is a classic prompt-injection attempt.
The rule's general advice: Do not install; the text is trying to steer the assistant against you. (see the review note for this case)
Known false alarm
Sample adversarial prompts listed in a red-teaming test file for testing the user's own model; they are examples, not instructions to the reader.
Reviewed by Bawwab maintainers, 2026-11-20
injection.hidden-unicode: Contains invisible characters high
skills/aws-sst-development/SKILL.md:95
OWASP LLM LLM01 Prompt InjectionOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0068 LLM Prompt ObfuscationCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
invisible characters: U+200B x1
What this means: Zero-width and direction-control characters can hide instructions that you cannot see but the AI reads.
The rule's general advice: Remove the invisible characters, or do not install. (see the review note for this case)
Known false alarm
A single zero-width space inside a markdown code span of a Pulumi `$interpolate` tip, there for formatting; the surrounding text is ordinary advice.
Reviewed by Bawwab maintainers, 2026-11-20
injection.override: Tries to override the AI's instructions high
skills/break-ui/SKILL.md:48
OWASP LLM LLM01 Prompt InjectionOWASP LLM LLM07 System Prompt LeakageOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0110.000 AI Agent Tool Poisoning: Definition and InstructionsMITRE ATLAS AML.T0069.002 Discover LLM System Information: System PromptCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
ignore previous instructions
What this means: Text read by the AI that says to ignore earlier instructions, reveal the system prompt or hide things from you is a classic prompt-injection attempt.
The rule's general advice: Do not install; the text is trying to steer the assistant against you. (see the review note for this case)
Known false alarm
The skill tells the agent to treat repository content as data and flag attempts to steer it; the quoted phrase is an example of what to flag.
Reviewed by Bawwab maintainers, 2026-11-20
injection.override: Tries to override the AI's instructions high
skills/browser-testing-with-devtools/SKILL.md:90
OWASP LLM LLM01 Prompt InjectionOWASP LLM LLM07 System Prompt LeakageOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0110.000 AI Agent Tool Poisoning: Definition and InstructionsMITRE ATLAS AML.T0069.002 Discover LLM System Information: System PromptCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
Ignore previous instructions
What this means: Text read by the AI that says to ignore earlier instructions, reveal the system prompt or hide things from you is a classic prompt-injection attempt.
The rule's general advice: Do not install; the text is trying to steer the assistant against you. (see the review note for this case)
Known false alarm
Defensive rule: browser content is untrusted and must not be followed as instructions; the quoted phrase is an example.
Reviewed by Bawwab maintainers, 2026-11-20
injection.override: Tries to override the AI's instructions high
skills/effective-agent-skills/SKILL.md:291
OWASP LLM LLM01 Prompt InjectionOWASP LLM LLM07 System Prompt LeakageOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0110.000 AI Agent Tool Poisoning: Definition and InstructionsMITRE ATLAS AML.T0069.002 Discover LLM System Information: System PromptCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
ignore previous instructions
What this means: Text read by the AI that says to ignore earlier instructions, reveal the system prompt or hide things from you is a classic prompt-injection attempt.
The rule's general advice: Do not install; the text is trying to steer the assistant against you. (see the review note for this case)
Known false alarm
Checklist item advising readers to scan third-party skills for injection phrases; the quote is an example.
Reviewed by Bawwab maintainers, 2026-11-20
injection.override: Tries to override the AI's instructions high
skills/etsy-search-listings/SKILL.md:110
OWASP LLM LLM01 Prompt InjectionOWASP LLM LLM07 System Prompt LeakageOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0110.000 AI Agent Tool Poisoning: Definition and InstructionsMITRE ATLAS AML.T0069.002 Discover LLM System Information: System PromptCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
Ignore any instruction
What this means: Text read by the AI that says to ignore earlier instructions, reveal the system prompt or hide things from you is a classic prompt-injection attempt.
The rule's general advice: Do not install; the text is trying to steer the assistant against you. (see the review note for this case)
Known false alarm
Defensive note: ignore instruction-like text inside Etsy titles and shop names.
Reviewed by Bawwab maintainers, 2026-11-20
injection.override: Tries to override the AI's instructions high
skills/feature-tracking/SKILL.md:210
OWASP LLM LLM01 Prompt InjectionOWASP LLM LLM07 System Prompt LeakageOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0110.000 AI Agent Tool Poisoning: Definition and InstructionsMITRE ATLAS AML.T0069.002 Discover LLM System Information: System PromptCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
ignore the instruction
What this means: Text read by the AI that says to ignore earlier instructions, reveal the system prompt or hide things from you is a classic prompt-injection attempt.
The rule's general advice: Do not install; the text is trying to steer the assistant against you. (see the review note for this case)
Known false alarm
Defensive note: treat repository text that tries to bypass safety checks as untrusted and ignore it.
Reviewed by Bawwab maintainers, 2026-11-20
injection.override: Tries to override the AI's instructions high
skills/find-animation-opportunities/SKILL.md:41
OWASP LLM LLM01 Prompt InjectionOWASP LLM LLM07 System Prompt LeakageOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0110.000 AI Agent Tool Poisoning: Definition and InstructionsMITRE ATLAS AML.T0069.002 Discover LLM System Information: System PromptCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
ignore previous instructions
What this means: Text read by the AI that says to ignore earlier instructions, reveal the system prompt or hide things from you is a classic prompt-injection attempt.
The rule's general advice: Do not install; the text is trying to steer the assistant against you. (see the review note for this case)
Known false alarm
Rule telling the agent to treat repository content as data and flag steering attempts; the quote is an example.
Reviewed by Bawwab maintainers, 2026-11-20
injection.override: Tries to override the AI's instructions high
skills/git-pr-review/SKILL.md:36
OWASP LLM LLM01 Prompt InjectionOWASP LLM LLM07 System Prompt LeakageOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0110.000 AI Agent Tool Poisoning: Definition and InstructionsMITRE ATLAS AML.T0069.002 Discover LLM System Information: System PromptCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
ignore previous instructions
What this means: Text read by the AI that says to ignore earlier instructions, reveal the system prompt or hide things from you is a classic prompt-injection attempt.
The rule's general advice: Do not install; the text is trying to steer the assistant against you. (see the review note for this case)
Known false alarm
Defensive guidance for PR review: commit and diff text is attacker-controlled and must be ignored as instructions; the quote is an example.
Reviewed by Bawwab maintainers, 2026-11-20
injection.html-comment: Hidden instruction in an HTML comment med
skills/hunt-deserialization/SKILL.md:53
OWASP LLM LLM01 Prompt InjectionOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0068 LLM Prompt ObfuscationCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
<!-- security-allowlist: documented payload decoding technique reference, do not execute outside authorized scope -->
What this means: HTML comments are invisible when a document is rendered but are still read by the AI, so they can carry instructions meant to stay unseen.
The rule's general advice: Read the comment; remove it or do not install. (see the review note for this case)
Expected for what this tool does
A `security-allowlist` comment after a documented attack example in an offensive-security reference skill. The repository uses these markers (164 of them, e.g. `security-allowlist: SEC005`) for its own content linter. The comment doesn't instruct the agent; the example itself is reference material, not something the skill runs.
Reviewed by Bawwab maintainers, 2026-11-20
injection.covert-action: Tells the AI to read or send sensitive data high
skills/hunt-dispatch/SKILL.md:119
OWASP LLM LLM01 Prompt InjectionOWASP LLM LLM02 Sensitive Information DisclosureOWASP Agentic ASI01 Agent Goal HijackOWASP Agentic ASI02 Tool Misuse and ExploitationMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0110.000 AI Agent Tool Poisoning: Definition and InstructionsMITRE ATLAS AML.T0086 Exfiltration via AI Agent Tool InvocationCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
.env → hunt-source-leak
What this means: Instructions to read keys or secrets, include them in replies or send them somewhere, or to do a hidden step first, are how injected text steals data.
The rule's general advice: Do not install unless you can explain why this text needs those files. (see the review note for this case)
Known false alarm
A routing table mapping observed signals (such as an exposed /.env) to the sibling skill that handles them; it does not tell the agent to read or send a file.
Reviewed by Bawwab maintainers, 2026-11-20
injection.html-comment: Hidden instruction in an HTML comment med
skills/hunt-k8s/SKILL.md:207
OWASP LLM LLM01 Prompt InjectionOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0068 LLM Prompt ObfuscationCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
<!-- security-allowlist: documented payload decoding technique reference, do not execute outside authorized scope -->
What this means: HTML comments are invisible when a document is rendered but are still read by the AI, so they can carry instructions meant to stay unseen.
The rule's general advice: Read the comment; remove it or do not install. (see the review note for this case)
Expected for what this tool does
A `security-allowlist` comment after a documented attack example in an offensive-security reference skill. The repository uses these markers (164 of them, e.g. `security-allowlist: SEC005`) for its own content linter. The comment doesn't instruct the agent; the example itself is reference material, not something the skill runs.
Reviewed by Bawwab maintainers, 2026-11-20
injection.override: Tries to override the AI's instructions high
skills/hunt-llm-ai/SKILL.md:60
OWASP LLM LLM01 Prompt InjectionOWASP LLM LLM07 System Prompt LeakageOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0110.000 AI Agent Tool Poisoning: Definition and InstructionsMITRE ATLAS AML.T0069.002 Discover LLM System Information: System PromptCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
Ignore previous instructions
What this means: Text read by the AI that says to ignore earlier instructions, reveal the system prompt or hide things from you is a classic prompt-injection attempt.
The rule's general advice: Do not install; the text is trying to steer the assistant against you. (see the review note for this case)
Known false alarm
Example prompt-injection strings listed as test payloads in a skill for assessing LLM apps; they are for the tester to send to the target, not instructions to the reader.
Reviewed by Bawwab maintainers, 2026-11-20
injection.html-comment: Hidden instruction in an HTML comment med
skills/hunt-rce/SKILL.md:104
OWASP LLM LLM01 Prompt InjectionOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0068 LLM Prompt ObfuscationCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
<!-- security-allowlist: documented payload technique reference, do not execute outside authorized scope -->
What this means: HTML comments are invisible when a document is rendered but are still read by the AI, so they can carry instructions meant to stay unseen.
The rule's general advice: Read the comment; remove it or do not install. (see the review note for this case)
Expected for what this tool does
A `security-allowlist` comment after a documented attack example in an offensive-security reference skill. The repository uses these markers (164 of them, e.g. `security-allowlist: SEC005`) for its own content linter. The comment doesn't instruct the agent; the example itself is reference material, not something the skill runs.
Reviewed by Bawwab maintainers, 2026-11-20
injection.covert-action: Tells the AI to read or send sensitive data high
skills/hunt-saml/SKILL.md:138
OWASP LLM LLM01 Prompt InjectionOWASP LLM LLM02 Sensitive Information DisclosureOWASP Agentic ASI01 Agent Goal HijackOWASP Agentic ASI02 Tool Misuse and ExploitationMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0110.000 AI Agent Tool Poisoning: Definition and InstructionsMITRE ATLAS AML.T0086 Exfiltration via AI Agent Tool InvocationCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
/etc/passwd contents leak
What this means: Instructions to read keys or secrets, include them in replies or send them somewhere, or to do a hidden step first, are how injected text steals data.
The rule's general advice: Do not install unless you can explain why this text needs those files. (see the review note for this case)
Expected for what this tool does
A chain note in a security-testing skill describing an XXE file-read payload as an example for authorized testing; it does not direct the agent to read files on the user's machine.
Reviewed by Bawwab maintainers, 2026-11-20
injection.html-comment: Hidden instruction in an HTML comment med
skills/hunt-saml/SKILL.md:84
OWASP LLM LLM01 Prompt InjectionOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0068 LLM Prompt ObfuscationCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
<!-- security-allowlist: documented payload decoding technique reference, do not execute outside authorized scope -->
What this means: HTML comments are invisible when a document is rendered but are still read by the AI, so they can carry instructions meant to stay unseen.
The rule's general advice: Read the comment; remove it or do not install. (see the review note for this case)
Expected for what this tool does
A `security-allowlist` comment after a documented attack example in an offensive-security reference skill. The repository uses these markers (164 of them, e.g. `security-allowlist: SEC005`) for its own content linter. The comment doesn't instruct the agent; the example itself is reference material, not something the skill runs.
Reviewed by Bawwab maintainers, 2026-11-20
injection.html-comment: Hidden instruction in an HTML comment med
skills/hunt-xss/SKILL.md:116
OWASP LLM LLM01 Prompt InjectionOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0068 LLM Prompt ObfuscationCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
<!-- security-allowlist: documented payload technique reference, do not execute outside authorized scope -->
What this means: HTML comments are invisible when a document is rendered but are still read by the AI, so they can carry instructions meant to stay unseen.
The rule's general advice: Read the comment; remove it or do not install. (see the review note for this case)
Expected for what this tool does
A `security-allowlist` comment after a documented attack example in an offensive-security reference skill. The repository uses these markers (164 of them, e.g. `security-allowlist: SEC005`) for its own content linter. The comment doesn't instruct the agent; the example itself is reference material, not something the skill runs.
Reviewed by Bawwab maintainers, 2026-11-20
injection.covert-action: Tells the AI to read or send sensitive data high
skills/hunt-xxe/SKILL.md:441
OWASP LLM LLM01 Prompt InjectionOWASP LLM LLM02 Sensitive Information DisclosureOWASP Agentic ASI01 Agent Goal HijackOWASP Agentic ASI02 Tool Misuse and ExploitationMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0110.000 AI Agent Tool Poisoning: Definition and InstructionsMITRE ATLAS AML.T0086 Exfiltration via AI Agent Tool InvocationCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
read of `/etc/passwd
What this means: Instructions to read keys or secrets, include them in replies or send them somewhere, or to do a hidden step first, are how injected text steals data.
The rule's general advice: Do not install unless you can explain why this text needs those files. (see the review note for this case)
Expected for what this tool does
A chain note in an XXE testing skill describing a /etc/passwd read through a target's XML parser; offensive-testing reference material, not an instruction to read local secrets.
Reviewed by Bawwab maintainers, 2026-11-20
injection.override: Tries to override the AI's instructions high
skills/improve-animations/SKILL.md:45
OWASP LLM LLM01 Prompt InjectionOWASP LLM LLM07 System Prompt LeakageOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0110.000 AI Agent Tool Poisoning: Definition and InstructionsMITRE ATLAS AML.T0069.002 Discover LLM System Information: System PromptCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
ignore previous instructions
What this means: Text read by the AI that says to ignore earlier instructions, reveal the system prompt or hide things from you is a classic prompt-injection attempt.
The rule's general advice: Do not install; the text is trying to steer the assistant against you. (see the review note for this case)
Known false alarm
Rule telling the agent to treat repository content as data and flag steering attempts; the quote is an example.
Reviewed by Bawwab maintainers, 2026-11-20
injection.override: Tries to override the AI's instructions high
skills/llm-security/SKILL.md:63
OWASP LLM LLM01 Prompt InjectionOWASP LLM LLM07 System Prompt LeakageOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0110.000 AI Agent Tool Poisoning: Definition and InstructionsMITRE ATLAS AML.T0069.002 Discover LLM System Information: System PromptCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
Ignore all previous instructions
What this means: Text read by the AI that says to ignore earlier instructions, reveal the system prompt or hide things from you is a classic prompt-injection attempt.
The rule's general advice: Do not install; the text is trying to steer the assistant against you. (see the review note for this case)
Known false alarm
An example attack prompt in a table of injection test levels for the user's own LLM app; it is not addressed to the reader.
Reviewed by Bawwab maintainers, 2026-11-20
injection.hidden-unicode: Contains invisible characters high
skills/llm-security/SKILL.md:73
OWASP LLM LLM01 Prompt InjectionOWASP Agentic ASI01 Agent Goal HijackMITRE ATLAS AML.T0051.001 LLM Prompt Injection: IndirectMITRE ATLAS AML.T0068 LLM Prompt ObfuscationCWE CWE-1427 Improper Neutralization of Input Used for LLM Prompting
invisible characters: U+200B x4
What this means: Zero-width and direction-control characters can hide instructions that you cannot see but the AI reads.
The rule's general advice: Remove the invisible characters, or do not install. (see the review note for this case)
Known false alarm
Four zero-width spaces inside a quoted example string that demonstrates a zero-width-character bypass technique in the same list of test examples.
Reviewed by Bawwab maintainers, 2026-11-20
10 low-severity notes (informational, do not affect the result)
files.binary-unscanned: Binary files not analysed low
5 binary files not analysed; use --deep to run in the sandbox
What this means: Prebuilt executables cannot be read by static rules, so what they do is unknown until they run.
What you can do: Use --deep to run the package in the sandbox, or prefer packages built from source.
install-dir-access: Reads or writes Claude's install folders low
skills/agent-orchestrator/scripts/scan_registry.py:47
MITRE ATLAS AML.T0081 Modify AI Agent Configuration
SKILLS_ROOT / ".claude" / "skills", # registered skills
What this means: Uses ~/.claude/agents, skills, commands or plugins (or a project .claude/ folder). Normal for installers, but it changes what Claude loads.
What you can do: Check which files it adds there and that you want them.
credential-access: Touches credentials or secrets low
skills/fact-check-x-unified/tests/trusted_search_config_test.py:55
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
previous = os.environ.copy()
What this means: Reads files or environment holding keys, tokens or sessions.
What you can do: Confirm the tool needs this; otherwise do not install.
credential-access: Touches credentials or secrets low
skills/hugging-face-paper-publisher/scripts/paper_manager.py:30
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
from dotenv import load_dotenv
What this means: Reads files or environment holding keys, tokens or sessions.
What you can do: Confirm the tool needs this; otherwise do not install.
install-dir-access: Reads or writes Claude's install folders low
skills/linkedin-content-generator/scripts/memory_manager.py:3
MITRE ATLAS AML.T0081 Modify AI Agent Configuration
Resolves memory.md relative to this script's location (inside .claude/skills/scripts/).
What this means: Uses ~/.claude/agents, skills, commands or plugins (or a project .claude/ folder). Normal for installers, but it changes what Claude loads.
What you can do: Check which files it adds there and that you want them.
install-dir-access: Reads or writes Claude's install folders low
skills/loki-mode/autonomy/run.sh:305
MITRE ATLAS AML.T0081 Modify AI Agent Configuration
"$HOME/.claude/skills/loki-mode/SKILL.md"
What this means: Uses ~/.claude/agents, skills, commands or plugins (or a project .claude/ folder). Normal for installers, but it changes what Claude loads.
What you can do: Check which files it adds there and that you want them.
install-dir-access: Reads or writes Claude's install folders low
skills/loki-mode/demo/record-full-demo.sh:142
MITRE ATLAS AML.T0081 Modify AI Agent Configuration
mkdir -p .claude/skills/loki-mode
What this means: Uses ~/.claude/agents, skills, commands or plugins (or a project .claude/ folder). Normal for installers, but it changes what Claude loads.
What you can do: Check which files it adds there and that you want them.
install-dir-access: Reads or writes Claude's install folders low
skills/obsess/scripts/install.py:106
MITRE ATLAS AML.T0081 Modify AI Agent Configuration
roots = (base / ".claude", base / ".config" / "opencode", base / ".agents")
What this means: Uses ~/.claude/agents, skills, commands or plugins (or a project .claude/ folder). Normal for installers, but it changes what Claude loads.
What you can do: Check which files it adds there and that you want them.
install-dir-access: Reads or writes Claude's install folders low
skills/skill-installer/scripts/install_skill.py:95
MITRE ATLAS AML.T0081 Modify AI Agent Configuration
CLAUDE_SKILLS = Path.cwd() / ".claude" / "skills"
What this means: Uses ~/.claude/agents, skills, commands or plugins (or a project .claude/ folder). Normal for installers, but it changes what Claude loads.
What you can do: Check which files it adds there and that you want them.
credential-access: Touches credentials or secrets low
skills/whatsapp-cloud-api/assets/boilerplate/nodejs/src/index.ts:2
OWASP LLM LLM03 Supply ChainOWASP Agentic ASI03 Identity and Privilege AbuseOWASP Agentic ASI04 Agentic Supply Chain VulnerabilitiesMITRE ATLAS AML.T0055 Unsecured CredentialsMITRE ATLAS AML.T0083 Credentials from AI Agent Configuration
import dotenv from 'dotenv';
What this means: Reads files or environment holding keys, tokens or sessions.
What you can do: Confirm the tool needs this; otherwise do not install.
Dispute this result
Maintainer and think this is wrong? Open a dispute on GitHub with the target id plugin-sickn33-agentic-awesome-skills-plugins-agentic-awesome-skills-claude-e9562a.