Caveat verdict
ai-company-unified
Consolidated AI company governance and department management templates with no concrete exfiltration, remote code execution, or manipulation behavior in the content despite the executionSinkDetected flag.
⚠ Flagged for review — coarse, uncorroborated signal, not a confirmed exploit. Review the config yourself before installing.
Automated static analysis — not a human review. Caveat flags capabilities, not confirmed intent, and can produce false positives. Disagree with this verdict? Use Dispute below.
Findings (14)
Possible hardcoded credential
references/data-integration.md · code · apiKey = 'sk_live_abc123xyz
Accesses .ssh directory
SKILL.md · frontmatter · .ssh/
Accesses cloud provider credentials
SKILL.md · frontmatter · ~/.aws
Instructs covert action — may act without user awareness
references/departments/intelligence.md · code · silently
Uses eval() — can execute arbitrary code
references/departments/intelligence.md · prose · downgraded · eval(
Recursive delete from root or home — destructive command
prompts/03-test-cases.md · prose · downgraded · rm -rf /
Accesses sensitive system files
prompts/03-test-cases.md · prose · downgraded · /etc/passwd
subprocess execution — runs system commands from Python
references/departments/platform-and-infrastructure.md · code · subprocess.run(
References agent memory files
README.md · code · memory.md
Uses exec() — may execute shell commands
references/departments/intelligence.md · prose · downgraded · exec(
Accesses sensitive environment variables
references/data-integration.md · code · process.env.TEST_API_KEY
Node http/https module — low-level network access
references/data-integration.md · code · require('https')
Popular HTTP library — network access
references/departments/quality-and-operations.md · code · got
Python os.environ.get — reads environment variable
references/integrations.md · code · os.environ.get(
Why the tier is capped
Execution sink present in raw bytes (Hard Floor: class D). Final tier capped at Caution — cannot be lifted by any downgrade, example-payload opt-in, or allowlist.
Permissions & capabilities
No declared permissions — minimal attack surface.
Is this flag fair?
Thanks — recorded.