The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The skill workflow requires the agent to execute local scripts (scripts/analyze_repo.py and scripts/generate_report.py). This is a standard and expected operation for the skill's primary function of repository analysis.
[EXTERNAL_DOWNLOADS] (SAFE): The README describes a workflow where the agent clones repositories to perform scans. This is the intended use case for the tool and is initiated by explicit user command.
[DATA_EXPOSURE & EXFILTRATION] (SAFE): Analysis of scripts/generate_report.py shows no hardcoded credentials, sensitive path access, or network calls. Data is processed locally and output to the terminal or a local file.
[INDIRECT_PROMPT_INJECTION] (LOW): The skill creates a vulnerability surface by ingesting and reporting on untrusted data from third-party repositories.
Ingestion points: scripts/analyze_repo.py (not provided) scans files like .eslintrc, pyproject.toml, and README.md.
Boundary markers: The markdown report does not use specific escaping or boundary markers for the 'reason' fields extracted from the analysis.
Capability inventory: The agent has the capability to write files (AGENTS.md, templates) and execute shell commands (linters, hooks) as part of the remediation workflow.
Sanitization: There is no evidence of sanitization for strings extracted from the scanned repository before they are presented to the agent in the report.
[NO_CODE] (SAFE): Although one script (analyze_repo.py) is missing from the provided files, the existing script and documentation clearly define a legitimate utility rather than malicious intent.

readiness-report