new-skill
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local project scripts (
fdk/tools/new-skill.py,harness/scripts/sync-skills.py, andharness/scripts/skill-registry.py) usingpython3. These commands are used to automate the creation of skill templates and maintain parity between the canonical skill tree and the llmwiki mirror. - [INDIRECT_PROMPT_INJECTION]: The skill accepts user-provided input for the skill name, loop type, and description, which are subsequently interpolated into shell commands. This creates a potential surface for command injection if the user provides malformed or malicious strings, although the skill includes a human-in-the-loop step by requiring the user to manually update certain registry files.
- Ingestion points: User-supplied values for the
<name>,--loop, and--descparameters inSKILL.mdsteps 1 and 2. - Boundary markers: No specific delimiters or "ignore instructions" markers are used for these input variables.
- Capability inventory: Execution of shell commands via
python3and file system write operations via the local scripts. - Sanitization: The instructions do not specify any validation or sanitization logic for the user-supplied strings before they are passed to the shell.
Audit Metadata