learning-aggregator-ci

Pass

Audited by Gen Agent Trust Hub on Oct 27, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill scans potentially untrusted data from .learnings/ files which are then summarized and posted as GitHub issue comments or used to trigger downstream workflows.
  • Ingestion points: Reads data from multiple files in the .learnings/ directory, including LEARNINGS.md, ERRORS.md, FEATURE_REQUESTS.md, and HEALS.md.
  • Boundary markers: The skill instructions do not specify any boundary markers or instructions to the agent to ignore embedded commands within the parsed files.
  • Capability inventory: The workflow can post issue comments (add-comment), upload artifacts (upload-artifact), and trigger downstream GitHub Actions workflows (call-workflow).
  • Sanitization: No explicit sanitization or filtering of the content extracted from the scanned files is defined before it is included in the output report.
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the installation of necessary external tools and components.
  • Fetches and installs the gh-aw extension from the official GitHub repository.
  • Downloads and adds the skill itself from the author's public repository (pskoett/pskoett-skills).
  • [COMMAND_EXECUTION]: Utilizes standard command-line tools for workflow management and validation.
  • Executes gh CLI commands and gh-aw extension utilities (gh aw compile) to build and validate the CI aggregation workflow.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 27, 2026, 09:18 PM