changelog-generator

Pass

Audited by Gen Agent Trust Hub on May 11, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes standard local Git commands like git log and git describe to retrieve repository metadata. These operations are read-only and occur within the local environment.
  • [PROMPT_INJECTION]: The skill processes commit messages, which represent untrusted external data from repository contributors. This creates an indirect prompt injection surface where a malicious actor could attempt to manipulate the LLM's output or formatting.
  • Ingestion points: Data is ingested through the output of git log commands as specified in SKILL.md.
  • Boundary markers: The instructions do not define protective delimiters or specific directions to ignore instructions embedded in commit messages.
  • Capability inventory: The skill uses the LLM for text summarization and categorization; it does not invoke any tools with file-writing or network capabilities.
  • Sanitization: There is no specified logic for validating or filtering the content of retrieved commit messages.
Audit Metadata
Risk Level
SAFE
Analyzed
May 11, 2026, 10:19 PM
Security Audit — agent-trust-hub — changelog-generator