The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external repositories to determine project conventions. This data includes issue titles, pull request titles, and repository template schemas fetched via the GitHub and Piper APIs.
Ingestion points: The OrientationGatherer class in lib/orient.dart fetches remote content using gh pr list, gh issue list, and gh api calls.
Boundary markers: The skill does not use specific delimiters or instructions to ignore potential injection within the orientation data, although it does require manual user approval before the final issue or pull request is submitted.
Capability inventory: The skill has the ability to create issues and pull requests via the gh CLI tool as defined in SKILL.md (Step 5) and lib/orient.dart.
Sanitization: No explicit sanitization or filtering of the fetched metadata (titles, labels, templates) is performed before it is presented to the agent.
[COMMAND_EXECUTION]: The skill utilizes the Process.run function in lib/orient.dart to execute system commands including gh (GitHub CLI) and p4 (Perforce/Piper). These commands are used to view repository status, list issues/PRs, and call GitHub APIs to gather project conventions.
[EXTERNAL_DOWNLOADS]: The skill fetches configuration and template data from GitHub's official API.
Metadata and issue form schemas are retrieved from repos/$remoteRepo/contents/ paths via the GitHub CLI's API tool.

github-post