Skills
Navigate
skills/fabioassuncao/issue-flow/init-repository/Gen Agent Trust Hub

init-repository

Pass

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources, including GitHub issues and local text documents, which could potentially contain malicious instructions.\n
  • Ingestion points: references/issue-input.md describes resolving issue content from GitHub URLs or local file paths.\n
  • Boundary markers: Instructions in references/repository-policy.md and references/issue-input.md emphasize treating input as "demand content" and explicitly state to inspect configuration as text without execution.\n
  • Capability inventory: The skill has the ability to write files to the repository via scripts/scaffold.mjs, perform Git operations, and execute the issue-flow CLI.\n
  • Sanitization: references/issue-input.md includes logic to reject invalid identifiers containing path traversal sequences like /, \\, or ...\n- [COMMAND_EXECUTION]: The skill executes local CLI tools to perform its intended tasks.\n
  • Evidence: scripts/optional-cli.mjs uses node:child_process.spawnSync to execute the issue-flow command.\n
  • Evidence: references/repository-policy.md details the use of the gh CLI for fetching labels and organization issue types.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 6, 2026, 05:52 AM