The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the camp CLI tool (e.g., camp workitem, camp wi) to interact with the local environment. It executes commands to query work items in JSON format, create new items, and adopt existing directories into the workflow tracking system.
[INDIRECT_PROMPT_INJECTION]: The skill processes content from various project directories (e.g., workflow/, intents, designs) to discover work items. This creates an attack surface where untrusted data embedded in those files could influence agent behavior during triage or planning tasks.
Ingestion points: Reads metadata and content from project-level workflow directories and documents via camp workitem --json (SKILL.md).
Boundary markers: No specific delimiters or instructions are provided to the agent to disregard instructions found within the discovered work items.
Capability inventory: The agent can execute shell commands (camp) and perform file/directory management based on the ingested data.
Sanitization: The skill instructions do not specify any validation or sanitization of the content retrieved from work items before processing.

camp-workitems