skills/jwh3times/skills/end-session/Gen Agent Trust Hub

end-session

Pass

Audited by Gen Agent Trust Hub on Nov 21, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources including Git commit messages, GitHub issue titles/bodies, and PR descriptions to generate session reports and update project boards.\n
  • Ingestion points: SKILL.md (Step 1, 3) reads from git log, gh issue list, and gh pr list.\n
  • Boundary markers: No explicit markers are used to delimit untrusted data in the generated output.\n
  • Capability inventory: Extensive use of git, gh CLI, and local file writes.\n
  • Sanitization: board.mjs uses JSON.stringify for GraphQL variables, which prevents CLI-level injection but does not mitigate the risk of the agent following malicious instructions embedded in issue text.\n- [DYNAMIC_EXECUTION]: The skill executes shell commands that are dynamically retrieved from repository settings files located in docs/agents/*.md.\n
  • Evidence: Settings such as Workspace status command, Sync with main, and Docs validation are extracted from markdown and executed as system commands.\n- [COMMAND_EXECUTION]: The skill relies on system commands (git, gh, node, bash) to manage the repository state and interact with GitHub APIs.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 21, 2026, 01:25 PM