higgsfield-assist
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions for the agent to execute a local Python utility script at
../../scripts/higgsfield_memory.py. This script is intended to calculate credit ratios and budgets for user projects. - [INDIRECT_PROMPT_INJECTION]: The skill documentation establishes an attack surface where the agent ingests data from external project manifests and local ledgers. 1. Ingestion points: Project names (
<project>) and manifest files (<manifest.json>) are passed as arguments to the execution script in SKILL.md. 2. Boundary markers: Not explicitly provided in the instructional text; the agent is directed to pass arguments directly to the shell. 3. Capability inventory: The skill is designed to perform command execution via subprocess to run the analytical scripts. 4. Sanitization: No explicit validation or escaping of the user-provided project name is described in the markdown, relying instead on the script implementation.
Audit Metadata