speckit-converge

Warn

Audited by Gen Agent Trust Hub on Aug 11, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a project-local shell script at .specify/scripts/bash/check-prerequisites.sh to initialize its context. This permits any code within that script to run in the agent's environment.- [REMOTE_CODE_EXECUTION]: The skill implements an extension hook mechanism that reads command definitions from .specify/extensions.yml. It dynamically constructs and automatically executes these commands if they are marked as mandatory, allowing project files to control agent behavior.- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes content from various project files to drive its assessment logic without sanitization.
  • Ingestion points: .specify/extensions.yml, spec.md, plan.md, tasks.md, and .specify/memory/constitution.md.
  • Boundary markers: Absent; the instructions do not include delimiters or warnings to ignore embedded instructions in these files.
  • Capability inventory: Execution of shell scripts, dynamic skill invocation via hooks, and appending content to tasks.md.
  • Sanitization: No escaping or validation of external markdown content is mentioned.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 11, 2026, 10:52 PM
Security Audit — agent-trust-hub — speckit-converge