Skills
Navigate
skills/samber/developer-relations-skills/oss-issue-triage/Gen Agent Trust Hub

oss-issue-triage

Pass

Audited by Gen Agent Trust Hub on Oct 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions include commands to export GitHub issue data using the gh CLI and process it with a local Python script. These operations are restricted to metadata extraction for project analysis and do not involve unauthorized system modifications.\n- [INDIRECT_PROMPT_INJECTION]: The skill defines an attack surface by ingesting external repository data (issues.json). Analysis of the evidence chain:\n
  • Ingestion points: The issues.json file generated in Step 1.\n
  • Boundary markers: Absent.\n
  • Capability inventory: Uses gh for extraction, python3 for processing, and writes state to devrel-context.md.\n
  • Sanitization: The analysis script (scripts/triage-baseline.py) uses standard json and csv parsers and only extracts metadata for statistical reporting (counts, medians) without executing or displaying untrusted content.\n- [SAFE]: The skill relies on a local script that uses only standard Python libraries, avoiding risks associated with external or unverified dependencies.\n- [SAFE]: All external references within the skill point to reputable open-source projects and established industry research.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 26, 2026, 08:31 PM