The Agent Skills Directory

[SAFE]: The skill instructions and associated templates were analyzed for security risks; no malicious patterns were identified.- [PROMPT_INJECTION]: The skill instructions define an interactive process for requirement gathering and automated creation of skill files. While this involves processing untrusted user data, no malicious injection or bypass patterns were found. Analysis of the indirect prompt injection surface follows: 1) Ingestion points: The skill reads user input and existing workspace files (SKILL.md, CLAUDE.md) to build context in Phase 2 and Phase 5; 2) Boundary markers: The instructions do not specify the use of delimiters for external data when generating output; 3) Capability inventory: The skill is capable of writing new files to the workspace as seen in Phase 5; 4) Sanitization: No explicit validation or escaping of the processed content is defined before file-write operations occur. This surface is inherent to the skill's utility and does not indicate malicious intent.- [DATA_EXFILTRATION]: No network-enabled tools or operations were identified. The skill's data access is limited to standard local context retrieval for planning purposes.- [COMMAND_EXECUTION]: The skill instructs the agent to save generated plans and skills to the workspace using standard file system tools, which is the intended functional behavior of the skill.

process-interviewer