The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external data which serves as a potential vector for indirect prompt injection.
Ingestion points: The skill reads PRD files located at projects/<project>/specs/prds/PRD-*.md and explores arbitrary modules within the codebase to identify vertical boundaries (Step 1, Step 2).
Boundary markers: Absent. There are no instructions provided to the agent to treat the content of the PRD or codebase as untrusted data or to use delimiters to separate these inputs from the skill's operational instructions.
Capability inventory: The agent has the capability to execute shell commands via the wiki CLI tool, which includes writing to the backlog, creating new specification files, and updating project indices (Step 5, Step 6, Step 9).
Sanitization: Absent. The skill uses content derived from the PRD to populate command arguments (such as <title>) and file contents (implementation and test plans) without defined validation or filtering steps.
[COMMAND_EXECUTION]: The skill frequently invokes a local CLI tool named wiki to manage the project lifecycle. These commands include create-issue-slice, forge plan, forge run, update-index, lint, and lint-semantic. While these appear to be project-specific management tools, their execution is driven by the interpretation of external PRD documents.

prd-to-slices