Self-Audit
Skill declares a 'self-audit' CLI script and 'audit/' directory that do not exist; undeclared sensitive config.json with LLM tool permissions is present but not documented.
Why this conclusion was reached
0/4 dimensions flaggedDeclared resources and inferred behavior are broadly aligned.
No obvious high-risk egress or execution signals were found.
There is no explicit malicious chain in the report.
Dependency information is incomplete, so supply-chain confidence stays limited.
What drove the risk score up
SKILL.md declares 'self-audit' as main CLI script but file does not exist
config.json with LLM system prompts and allowed-tools mapping is not mentioned in documentation
Skill has zero executable code despite claiming CLI functionality
Most important evidence
Declared entry point does not exist
SKILL.md lists 'self-audit' as the 'Main CLI script' but no such file exists in the package. This is a significant doc-to-code mismatch.
SKILL.md:66 Undeclared sensitive configuration file
config.json contains LLM system prompts and tool permission mappings but is not mentioned anywhere in SKILL.md. This file includes allowed-tools definitions (Bash, Read, Write, WebFetch) which map to resource permissions.
config.json:1 Declared storage directory does not exist
SKILL.md mentions an 'audit/' directory for audit log storage that does not exist.
SKILL.md:67 Declared capability vs actual capability
No scripts exist to infer capabilities No network code exists No shell scripts exist No env access code exists Suspicious artifacts and egress
No obvious IOC was extracted.
Dependencies and supply chain
There are no structured dependency warnings.
File composition
config.json SKILL.md