The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's workflow involves reading existing documents from the project environment to automatically extract context for new PRDs. This creates a surface for indirect prompt injection where malicious instructions embedded in those documents could potentially influence the agent's actions during the extraction and generation phases.
Ingestion points: The skill searches for and reads files in docs/ or prd/ directories, as well as any user-specified files (SKILL.md, Phase 0).
Boundary markers: There are no instructions for the agent to use delimiters or ignore instructions found within these ingested files.
Capability inventory: The skill has the capability to write files to the local system and execute the lark-cli tool to interact with cloud documents (SKILL.md, Phase 4).
Sanitization: No sanitization or validation logic is defined for the content extracted from external files before it is used in the prompt context.
[COMMAND_EXECUTION]: The skill relies on the execution of the lark-cli command-line interface to perform its primary function of uploading documents to Feishu. This involves interpolating generated content into shell commands (SKILL.md, Phase 4).

create-prd