The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of local CLI tools canon and orchestra to interact with a SQLite database and manage git worktrees. These operations are core to the skill's intended functionality for workflow automation.
[DYNAMIC_EXECUTION]: The skill executes shell commands defined in verify_command fields of acceptance records via orchestra check. While this allows execution of logic not contained within the skill itself, it is the documented method for running tests and verification scripts within the system.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from a local store (records) which influences its decision-making (e.g., assessing if a requirement is 'verified').
Ingestion points: Records are read from the local SQLite store via canon list, orchestra view, and orchestra brief commands.
Boundary markers: The system uses structured data (JSON/NDJSON) and explicit schemas for all record exchanges.
Capability inventory: The skill can execute shell commands via orchestra check, update records via canon apply, and commit changes via git.
Sanitization: The skill relies on canon validate to enforce schema compliance and uses revision tracking (--expected-revision) to prevent unauthorized or concurrent modifications.

verify