dev-prompt-canary

Pass

Audited by Gen Agent Trust Hub on Nov 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the execution of shell commands to set up a development environment, including git worktree add for isolation and ln -s for symlinking local resources such as node_modules and .env files.
  • [DYNAMIC_EXECUTION]: The validation process involves generating code files (grader.ts) and executing them at runtime via runPlaywright to verify the accuracy of the grading logic against specific solution patches.
  • [INDIRECT_PROMPT_INJECTION]: The skill identifies a vulnerability surface where the agent ingests external data from Markdown files (such as task.md and expectations.md) to drive the generation of code and prompts. 1. Ingestion points: guides/, task.md, expectations.md (found in SKILL.md). 2. Boundary markers: None explicitly defined in instructions. 3. Capability inventory: Shell command execution, code generation, and test execution (runPlaywright) mentioned in SKILL.md. 4. Sanitization: Includes static AST linting using src/core/grader-lint.ts to check for violations in generated code.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 10, 2026, 09:41 AM