The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute specific local shell scripts (check_paths.sh and check_progress.sh) during the verification phase. These scripts are used to validate the state of the repository relative to the development plan.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources such as development plans, PRDs, and review reports.
Ingestion points: These documents are read at runtime as specified in SKILL.md (e.g., locate_requested_plan, review_report_source).
Boundary markers: The skill does not define explicit delimiters for this external content, relying instead on structural section identification within the plans.
Capability inventory: The skill has capabilities for file system writes (modifying plans, creating decision records and CRD files), local script execution, and communication via user-interaction tools.
Sanitization: The skill mitigates injection risks by implementing a 'verify-before-accept' policy (e.g., RULE @evidence_and_collaboration) and mandating that all critical decisions and external handoffs must be explicitly confirmed by the user through a blocking interaction tool.

resolve-blocked