The Agent Skills Directory

[SAFE]: The skill operates safely within its intended purpose of managing planning artifacts. It is explicitly prohibited from modifying implementation code.\n- [COMMAND_EXECUTION]: The skill uses the openspec CLI for operations like store listing and status checks. Access to the shell is restricted by the allowed-tools configuration to only the openspec command.\n- [INDIRECT_PROMPT_INJECTION]: The skill contains an ingestion surface by reading Markdown artifacts from existingOutputPaths. No explicit boundary markers or sanitization methods are defined. However, the associated capabilities are limited to the openspec command set via Bash, and the skill is strictly blocked from implementation-level edits, maintaining a safe risk profile.

openspec-update-change