gsd-debug
Pass
Audited by Gen Agent Trust Hub on Aug 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Executes local Node.js utility scripts (
gsd-tools.cjs) located in the vendor's internal toolkit directory ($HOME/.codex/get-shit-done/bin/) to load session state and resolve model configurations. - [COMMAND_EXECUTION]: Performs file system operations using standard shell utilities (
ls,grep,cat) to manage persistent debug session files within the.planning/debug/directory. - [PROMPT_INJECTION]: The skill maintains an indirect prompt injection surface by interpolating untrusted user input into the instructions provided to spawned subagents.
- Ingestion points: User input captured via
{{GSD_ARGS}}and interactiveAskUserQuestionworkflows inSKILL.md. - Boundary markers: Employs XML-style delimiters (
<objective>,<symptoms>,<mode>) to separate user-provided data from agent instructions. - Capability inventory: The orchestrator has the ability to execute shell commands, read local files, and spawn subagents (
spawn_agent). - Sanitization: No explicit sanitization or escaping mechanisms are applied to the symptom data before it is passed to the subagent's context.
Audit Metadata