The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the openspec CLI to automate workflow tasks. It runs commands like openspec status and openspec instructions to manage state. These are standard operations for this tool's intended use case.\n- [INDIRECT_PROMPT_INJECTION]: The skill reads context and operationGuidance from external CLI output, which acts as a potential surface for indirect instructions.\n
Ingestion points: Data returned by openspec instructions apply --json as described in SKILL.md.\n
Boundary markers: The instructions include specific guardrails requiring the agent to prioritize built-in workflow logic and report any conflicts found in the external guidance.\n
Capability inventory: The agent can read project context files and execute commands within the openspec tool scope.\n
Sanitization: The skill mandates a logical review process where external guidance is evaluated against known project facts and constraints before being followed.

openspec-apply-change