starting-an-implementation-plan
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute various shell commands to manage the development environment.
- Evidence: Use of
git rev-parse --show-toplevel,ls -d,git checkout -b, andcatto verify paths and restore instruction context. - The skill uses a 'slug' extracted from a filename and user-provided branch names directly in shell commands, which presents a potential command injection risk if the agent does not perform adequate sanitization.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external files which could contain malicious instructions designed to influence the agent's behavior during the implementation process.
- Ingestion points: Reads design plans from
docs/design-plans/and implementation guidance from.ed3d/implementation-plan-guidance.md(SKILL.md). - Boundary markers: No explicit delimiters or instructions to ignore embedded commands are specified for these ingested files.
- Capability inventory: The skill can create and update tasks (
TaskCreate,TaskUpdate), execute shell commands (git,ls,cat), and invoke other planning/execution sub-skills. - Sanitization: There is no evidence of sanitization or filtering of the content read from the design plans or guidance files before it is incorporated into the agent's context.
Audit Metadata