speckit-auto

Pass

Audited by Gen Agent Trust Hub on Nov 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONCREDENTIALS_UNSAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches the specify-cli from GitHub and the superpowers library from its respective repository during provider installation or recovery. These downloads are pinned to specific, immutable commits and are required for the skill's primary functionality.
  • [COMMAND_EXECUTION]: Extensive use of shell commands for git worktree management, submodule initialization, and environment-scrubbed application verification. The skill also executes a provided Node.js script for structured event logging.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from Jira tickets and repository files. It includes a specific safety protocol (Rule 12) that explicitly instructs the agent to ignore instructions embedded in this data that attempt to alter the pipeline's execution or access restricted resources.
  • [CREDENTIALS_UNSAFE]: The skill handles Jira API tokens for ticket intake. It follows best practices by using curl configuration files via stdin to prevent exposing secrets in process lists and implements a redaction engine to strip recognizable secret patterns (GitHub tokens, AWS keys, etc.) from its own event logs.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 17, 2026, 02:00 PM