refactor-archeology

Pass

Audited by Gen Agent Trust Hub on Feb 28, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to run standard git commands like git blame and git log to analyze repository history. These are routine operations for the stated purpose of code archaeology.
  • [DATA_EXPOSURE]: The skill accesses local git metadata, commit messages, and source code comments. This is restricted to local repository data and no network exfiltration patterns were detected.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data such as commit messages and pull request descriptions which could theoretically contain malicious instructions.
  • Ingestion points: Git commit messages, PR descriptions, and code comments are read from the local repository (SKILL.md).
  • Boundary markers: Not explicitly defined in the instructions.
  • Capability inventory: Limited to read-only git operations and text analysis; no file-write, network, or arbitrary code execution capabilities are defined in the skill itself.
  • Sanitization: Not explicitly defined, though the risk is low as the output is a research report.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 28, 2026, 11:34 PM