The Agent Skills Directory

[SAFE]: The skill provides a robust framework for managing technical documentation within a repository using predefined templates for Specs, Plans, and Decisions.
[EXTERNAL_DOWNLOADS]: The skill references external GitHub repositories (Fission-AI and Matt Pocock) for schema definitions and interview methodologies. These are used for documentation and attribution purposes rather than execution, and are well-known developer resources.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface as it reads existing project documentation and source code to generate plans. While this is an inherent risk for agents processing repository content, the skill mitigates this by using highly structured templates and requiring explicit user confirmation during the 'Grill' and 'Propose' phases.
[COMMAND_EXECUTION]: The operations described involve standard filesystem and version control actions (e.g., git mv, file writing) that are necessary for the skill's stated purpose of managing repository documents.
[PERSISTENCE]: The skill modifies project-specific instruction files (CLAUDE.md or AGENTS.md) to include documentation conventions. This is a legitimate configuration pattern for repository-aware agents to maintain context across sessions and does not constitute a malicious persistence mechanism.

spec-flow