Skills
Navigate
skills/pskoett/pskoett-ai-skills/learning-aggregator-ci/Gen Agent Trust Hub

learning-aggregator-ci

Pass

Audited by Gen Agent Trust Hub on Jun 30, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted data from repository files (.learnings/*.md) and interpolates it into GitHub issue comments and YAML artifacts. This creates a surface for indirect prompt injection where malicious content in the learning files could influence the generated report or downstream workflows like eval-creator-ci.
  • Ingestion points: The workflow parses LEARNINGS.md, ERRORS.md, FEATURE_REQUESTS.md, and HEALS.md from the .learnings/ directory.
  • Boundary markers: No explicit boundary markers or 'ignore' instructions are used when reading the markdown files.
  • Capability inventory: The skill uses add-comment to post results, upload-artifact to persist data, and call-workflow to trigger secondary processes.
  • Sanitization: The instructions do not specify any sanitization, escaping, or validation of the text content extracted from the learning files before it is included in the outputs.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 30, 2026, 06:59 AM
Security Audit — agent-trust-hub — learning-aggregator-ci