The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the ## Reconciliation Brief section of review-NNN.md to determine which changes to apply and which commands to run.
Ingestion points: Processes review-NNN.md and review-NNN.toml files.
Boundary markers: Relies on markdown headers (e.g., ## Reconciliation Brief) to isolate data, but lacks strict sanitization to prevent the agent from following instructions embedded within the brief itself.
Capability inventory: Has the ability to perform direct file writes to project artifacts (design.md, slice-NNN.md) and execute a variety of shell commands through the doctrine CLI.
Sanitization: No explicit sanitization, escaping, or validation of the brief's content is described before it is used to interpolate command arguments or file content.
[COMMAND_EXECUTION]: The skill is instructed to execute several doctrine CLI verbs (e.g., doctrine revision new, doctrine revision change add). These commands use parameters such as slugs, target IDs, and action types that are derived from the reconciliation brief, representing a standard command execution surface for the agent.

reconcile