The Agent Skills Directory

[SAFE]: The skill's primary function is to automate the collection of task outcomes and update project logs. Its behavior is consistent with its stated purpose and does not exhibit malicious intent.
[COMMAND_EXECUTION]: The skill utilizes several CLI commands (doctrine slice list, doctrine slice notes, doctrine check gate) to manage project data. These commands are executed within the context of the workspace to facilitate documentation tasks.
[PROMPT_INJECTION]: The skill incorporates a surface for indirect prompt injection by reading and following instructions from workspace files like harvest.md and notes.md. While this creates an ingestion point for untrusted data, it is a functional requirement for the skill's documentation logic.
Ingestion points: harvest.md, notes.md, and project state files (state/.../phases/).
Boundary markers: None identified; the agent is instructed to follow the procedure defined in the external files directly.
Capability inventory: Execution of doctrine CLI tools and file-write access to the local workspace.
Sanitization: No specific validation or escaping of the content read from the workspace files is performed.

harvest