autoimplement
Pass
Audited by Gen Agent Trust Hub on Aug 6, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes implementation plans which are untrusted external markdown files, creating a surface for indirect prompt injection.
- Ingestion points: Plan files are read from user-specified or convention-based paths (SKILL.md).
- Boundary markers: The generator subagent prompt uses explicit <PHASE_CONTENT> delimiters and negative constraints to isolate instructions (SKILL.md).
- Capability inventory: The skill dispatches subagents with file-write and command-execution capabilities, and the orchestrator can call other agent tools.
- Sanitization: The skill implements automated static analysis (Check 4) to block writes to sensitive paths like .env or .ssh, and enforces multi-model review chains for both input plans and output diffs.
- [COMMAND_EXECUTION]: The skill utilizes git, grep, and other shell utilities for workspace state management, validation of plan structure, and metadata tagging. These operations are consistent with the skill's automation purpose.
Audit Metadata