The Agent Skills Directory

[SAFE]: No security issues were detected during the analysis. The skill implements a comprehensive planning process with the following characteristics:
[PROMPT_INJECTION]: No override markers, bypass attempts, or instructions to ignore safety guidelines were found. The instructional language is strictly professional and task-oriented.
[DATA_EXPOSURE]: The skill operates on project-specific documentation within docs/specs/ and docs/plans/. It does not attempt to access sensitive system files, environment variables, or credentials. No hardcoded secrets are present.
[OBFUSCATION]: No Base64 encoding, zero-width characters, homoglyphs, or other obfuscation techniques were detected in the instructions or templates.
[REMOTE_CODE_EXECUTION]: There are no patterns involving remote script downloads or execution. The skill mentions well-known testing frameworks like fast-check and Hypothesis, which are standard tools for its stated purpose.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-defined specification files. While this is an ingestion point for untrusted data, the capabilities are limited to workflow transitions and task management within the Exarchos system, presenting no significant risk vector.
[DYNAMIC_CONTEXT_INJECTION]: No shell execution patterns using the !command syntax were found in the SKILL.md file.

plan