night-market-research-methodology

Pass

Audited by Gen Agent Trust Hub on Jul 31, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides functional shell commands for repository maintenance, such as scanning for dangling documentation references using rg, sed, and git check-ignore.
  • [COMMAND_EXECUTION]: Instructions include commands for running Python-based test suites and mutation testing using uv, pytest, and mutmut. These are standard development practices for ensuring code quality and test coverage.
  • [EXTERNAL_DOWNLOADS]: The skill references the installation of the mutmut package from a standard registry. This tool is a well-regarded utility for mutation testing in the Python ecosystem.
  • [COMMAND_EXECUTION]: The skill utilizes the GitHub CLI (gh) and the GraphQL API to retrieve discussion data from the author's own repository (athola/claude-night-market). This is used for tracking project history and decision-making.
  • [PROMPT_INJECTION]: The methodology involves ingesting external research data. The skill addresses potential indirect prompt injection risks by prescribing a structured synthesis process, multi-gate validation (the 'Evidence Bar'), and mandatory adversarial refutation before research findings are integrated into the repository.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 31, 2026, 01:38 AM
Security Audit — agent-trust-hub — night-market-research-methodology