The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from PR bodies, comments, review threads, and linked issues to reconstruct execution plans during the 'adoption' phase.
Ingestion points: Data enters context in references/adopt-pr.md via list-issue-comments, get-pr, list-review-comments, and get-issue operations.
Boundary markers: The skill explicitly defines an 'Untrusted content boundary' in references/agentic-setup.md, instructing the agent to treat repo/tracker content as data rather than instructions and to quote suspected injections rather than following them.
Capability inventory: The skill uses git commands for worktree management and commits, shell commands for validation gates, and tracker operations for PR mutations (SKILL.md, references/worktree-setup.md).
Sanitization: External values (IDs, slugs, branch names) are validated against numeric or alphanumeric regex patterns before interpolation.
[COMMAND_EXECUTION]: The skill executes shell commands for repository management, dependency installation, and validation gates.
Evidence: references/worktree-setup.md contains bash snippets for git worktree operations. SKILL.md step 6 executes validation.commands defined in the repository configuration.
Risk: These are standard operations for a development agent and are gated by project-local configuration.

om-auto-continue-pr