writing-plans
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes
git rev-parse HEADto capture the repository's current state for inclusion in the implementation plan. This is a standard, read-only operation used for project documentation. - [INDIRECT_PROMPT_INJECTION]: This skill functions as a plan generator for multi-step tasks, where the resulting markdown file is intended to be consumed by subsequent agents or sub-agents (referenced as
superpowers:executing-plans). This creates a multi-step pipeline where the plan's instructions influence later agent behavior. The skill mitigates risks associated with this surface by requiring strict formatting, mandatory impact analysis, and a Test-Driven Development (TDD) structure for all tasks. - [DATA_EXPOSURE]: The skill instructs the agent to modify
.git/info/excludeto ensure that generated plan files are not committed to the repository. Although this involves modifying repository metadata, it is a legitimate use of local Git configuration to manage private or temporary project artifacts.
Audit Metadata