The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill instructions in references/workflow-archive.md and references/workflow-save.md guide the agent to construct shell commands containing user-controlled data, such as requirement fields and node names, without explicit escaping or sanitization. This pattern, particularly when using inline strings for JSON or names in CLI arguments, creates a risk of command injection if the input contains shell-metacharacters.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data from text, tickets, and screenshots to drive analysis and archival tasks. Ingestion points: Data enters via user text, ticket descriptions/remarks, and multimodal screenshots as defined in references/workflow-step1-material.md. Boundary markers: The workflow does not include explicit delimiters or instructions to ignore potential commands embedded within the analyzed content. Capability inventory: The skill uses the cawplan CLI to perform write operations, including creating and updating requirements and module tree nodes. Sanitization: The skill only performs basic HTML stripping on ticket remarks, which is insufficient to prevent sophisticated indirect injection attacks.

cawplan-requirement-analyze