The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data from external sources such as tracker issues and repository content, which creates a potential injection surface.
Ingestion points: Issue metadata and bodies are fetched via the get-issue operation in SKILL.md (Step 1), and pull request diffs are accessed via get-pr-diff in SKILL.md (Step 10).
Boundary markers: The skill includes a dedicated 'Untrusted content boundary' in references/agentic-setup.md which explicitly instructs the agent to treat repository and tracker content strictly as data and to refuse compliance with any embedded directives (e.g., 'ignore previous instructions').
Capability inventory: The skill executes shell commands for worktree and branch management and invokes multiple sub-skills to perform implementation and verification tasks.
Sanitization: references/worktree-setup.md and references/agentic-setup.md mandate that issueId be purely numeric and that branch slugs be generated via a restrictive lowercase alphanumeric sanitation process before interpolation into shell commands.
[COMMAND_EXECUTION]: The skill performs shell operations to manage an isolated development environment.
Evidence: Shell scripts in references/worktree-setup.md perform git worktree add, git checkout, and git worktree remove operations.
Context: These operations are conducted within an isolated temporary directory (.ai/tmp/om-auto-fix-issue) to protect the user's primary repository state and include specific requirements for redacting secrets from all output and logs.

om-auto-fix-issue