The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from git logs and commit messages to verify claims, which represents a theoretical injection surface.
Ingestion points: The agent parses output from git show, git log, and git diff to extract concrete claims as described in SKILL.md.
Boundary markers: None; the skill relies on the agent's ability to selectively extract specific text patterns like "modifies X" or "adds Y".
Capability inventory: The skill uses Bash to execute git commands in SKILL.md.
Sanitization: None; the agent evaluates raw commit log text.
[EXTERNAL_DOWNLOADS]: The skill references authoritative documentation from well-known technology services.
Evidence: Links to git-scm.com (official Git site) and code.claude.com (Anthropic developer portal) provided in references/official-docs.md.
[SAFE]: The skill implements standard developer safety checks to prevent common errors like pushing the wrong branch or losing work during rebases. No malicious code, obfuscation, or unauthorized access patterns were found.

pr-diff-verification