update-skillranking-readme

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Executes a multi-line Python script using a shell heredoc to connect to a database and write updates to the local README.md file.
  • [CREDENTIALS_UNSAFE]: Accesses a sensitive configuration file (.env) located at a hardcoded absolute local path (/Users/longsen/Documents/code/aiworkskills/clawhub/.env). This file is used to retrieve the DATABASE_URL and SSL settings.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted data from a database and writes it to a public markdown file without robust escaping. 1. Ingestion points: SQL query results from the clawhub_skills table. 2. Boundary markers: None used when interpolating data into README.md. 3. Capability inventory: Local file writing (README.md) and shell command execution. 4. Sanitization: Minimal sanitization consisting of newline removal and string truncation to 300 characters.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 05:21 AM
Security Audit — agent-trust-hub — update-skillranking-readme