github-commit-recovery

Pass

Audited by Gen Agent Trust Hub on Jun 7, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill utilizes curl, git fetch, and a Python script with the requests library to download remote content from GitHub. These operations target a well-known service (GitHub) and are consistent with the forensic intent of the skill.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted commit data from external repositories, creating an attack surface for indirect prompt injection where malicious instructions could be embedded in commit diffs or metadata.
  • Ingestion points: Commit data and patches fetched from api.github.com and github.com (SKILL.md).
  • Boundary markers: Absent; there are no instructions for the agent to treat retrieved commit content as untrusted data or to ignore embedded instructions.
  • Capability inventory: Includes network operations (curl, requests) and local file system writes (Python snippet in SKILL.md).
  • Sanitization: Absent; retrieved content is processed or stored without validation or escaping.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 7, 2026, 01:29 PM