The Agent Skills Directory

[SAFE]: The skill's primary function is to orchestrate release workflows. All identified scripts (preflight_checker.py, changelog_generator.py, version_bumper.py, release_readiness_scorer.py) perform standard repository management tasks.
[COMMAND_EXECUTION]: The skill uses subprocess.run to execute standard development tools such as git, pytest, npm audit, cargo audit, and various linters (ruff, eslint, clippy). These operations are necessary for the skill's stated purpose and are implemented using best practices (e.g., passing arguments as lists to prevent shell injection).
[DATA_EXFILTRATION]: While preflight_checker.py contains regex patterns for sensitive credentials (AWS keys, GitHub tokens, etc.), these are used exclusively for local scanning to alert the user of potential secrets in their codebase. No network operations were found that would exfiltrate this or any other data.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from git logs and project manifests (e.g., package.json). However, it processes this data using deterministic parsers (regex and JSON/TOML libraries) rather than interpolating it directly into agent instructions.
Ingestion points: Git commit messages (parsed by changelog_generator.py and version_bumper.py), project manifest files like package.json, pyproject.toml, etc.
Boundary markers: None (uses regex-based structured parsing).
Capability inventory: File system writes (to update version files), CLI execution (git, linters, test runners).
Sanitization: Standard regex parsing and JSON/TOML decoding.

release-orchestrator