The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local system commands including openspec, git log, and mkdir. These are used to verify the environment, analyze recent project activity, and initialize the local directory structure required for the tool's operation. These actions are standard for developer workflow utilities and are executed using parameters derived from the local environment.- [PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection in Phase 2: Task Selection, where it ingests untrusted data by scanning the user's codebase for comments such as TODO, FIXME, and HACK. This data is then used to suggest and implement tasks.
Ingestion points: Local codebase scanning for comments and code patterns in Phase 2.
Boundary markers: None; code comments are parsed directly into the task selection logic.
Capability inventory: The skill has the ability to read and modify the local filesystem and execute the openspec CLI tool.
Sanitization: No explicit sanitization or filtering of scanned code content is performed before it is used to influence the agent's implementation tasks.

openspec-onboard