implement-plan
Pass
Audited by Gen Agent Trust Hub on May 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run automated verification checks, specifically mentioning
make check test. This executes commands defined in the local environment's Makefile. - [INDIRECT_PROMPT_INJECTION]: The skill processes instructions from external files located in
thoughts/shared/plans/and implements changes based on their content. - Ingestion points: Technical plans are read from the
thoughts/shared/plans/directory. - Boundary markers: There are no explicit instructions or delimiters used to separate the plan's instructions from the agent's core system prompt or to warn the agent to ignore potentially malicious instructions within those plans.
- Capability inventory: The skill allows the agent to read arbitrary files, modify the plan files (Edit capability), and execute shell commands (e.g.,
make). - Sanitization: The skill lacks content validation or sanitization for the plans it processes, relying on the user to provide safe plans.
Audit Metadata