skills-eval-triage

Pass

Audited by Gen Agent Trust Hub on Sep 10, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external evaluation reports (HTML and JSON formats) and uses this information to guide the agent in modifying the codebase (SKILL.md and scripts).
  • Ingestion points: The skill reads evaluation reports from tools/ci/gitlab/skills/report/ and logs from _logs/.
  • Boundary markers: There are no explicit delimiters or instructions to ignore malicious content embedded within the reports being triaged.
  • Capability inventory: The skill has the capability to execute shell scripts (update_scorecard.py, consolidate_reports.py, validate-skills-local.sh) and write modifications to files in the repository.
  • Sanitization: The skill instructions do not specify sanitization or validation steps for the content parsed from the reports before it is used to influence code changes.
  • [COMMAND_EXECUTION]: The skill instructions require the execution of various local Python and shell scripts to manage reports and verify fixes.
  • Evidence: Instructions include running python3 tools/ci/gitlab/skills/update_scorecard.py, python3 tools/ci/gitlab/skills/consolidate_reports.py, and tools/ci/gitlab/skills/validate-skills-local.sh.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 10, 2026, 02:57 PM
Security Audit — agent-trust-hub — skills-eval-triage