/aitask-audit-wrappers
Audit and port aitask skill wrappers across code-agent trees, plus helper-script whitelist coverage
Framework Development Skills
Code-agent skills used to develop the aitasks framework itself, not by end users
This subsection documents skills that exist in .claude/skills/ but are oriented toward people developing the aitasks framework — adding new skills, registering new code-agent models, auditing skill wrappers across agent trees. End users adopting the framework via the install tarball typically do not invoke them.
These pages live under development/ (rather than the user-facing Skills overview) so the user-facing skill table stays focused on day-to-day workflows.
Pure framework-development skills
These skills exist solely to support framework development. End users adopting aitasks via the install tarball typically do not invoke them.
| Skill | Description |
|---|---|
/aitask-audit-wrappers | Audit and port aitask skill wrappers across all four code-agent trees, plus helper-script whitelist coverage |
Useful for framework development and normal use
These skills live under Skills (their canonical location) but are surfaced here too because they support framework-development workflows — registering new code-agent models, refreshing model registries, generating release changelogs.
| Skill | Description |
|---|---|
/aitask-add-model | Register a known code-agent model in models_<agent>.json and optionally promote it to default |
/aitask-refresh-code-models | Research latest AI code agent models and update model configuration files |
/aitask-changelog | Generate changelog entries from commits and archived plans |
See also
- User-facing skills: Skills overview
- Framework architecture: Development Guide
- Skill authoring conventions: see “WORKING ON SKILLS / CUSTOM COMMANDS” in CLAUDE.md