spec-workflow-init
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONPERSISTENCEINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses standard shell utilities such as
git,ls,grep,find, andcatduring the initialization process to detect project properties like the package manager, programming language, and branch names. These operations are restricted to local environment discovery. - [PERSISTENCE]: The skill includes an optional step to register a
Stophook in.claude/settings.json. This hook is designed to execute a local script (pipeline-watchdog.sh) to prevent silent stalls in the development pipeline. This modification requires explicit user approval through an interactive prompt. - [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection because it reads project files like
package.jsonto generate instructions for sub-agents. However, this risk is mitigated by an environment detection summary presented to the user for validation before any files are generated, and the resulting instructions are human-readable markdown files intended for review.
Audit Metadata