pskoett-ai-skills plugin
The plugin connects task framing, code checks and saved lessons so agents can use past mistakes in later sessions.
- 332 ★ on the repository
- Last commit
Extended access
Why this level
- The plugin adds hooks, stores local lessons and runs project checks; persistent instructions change after human approval.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/pskoett-ai-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The plugin adds hooks, stores local lessons and runs project checks; persistent instructions change after human approval.
This is third-party code. Review the repository files before installing.
What it does
pskoett-ai-skills provides coordinated agent workflows. The agent frames the goal, retrieves useful session evidence, watches for scope drift and runs project checks. On failure, it diagnoses, fixes and records the result for reuse. In a separate cycle, the agent groups recurring lessons and proposes project-rule changes. Human approval precedes instruction updates, after which regression checks can help catch recurrence.
Who it is for. Technical leads using coding agents who want repeated team mistakes to become testable project guidance.
Good fit when
- An agent repeats a mistake fixed in a previous session.
- A long conversation drifts from the original goal.
- Implementation needs project checks and a separate quality review.
- Accumulated lessons need selection for future project guidance.
Not a fit when
- You expect every client to run hooks from the portable package automatically.
- Session history must not be read through available services.
- Project rules should be rewritten without a responsible person’s approval.
Example request
Investigate this recurring test failure, record its cause and propose a project rule for review before applying it.Limitations
Client packages differ: the portable version contains skills, while hooks use native adapters. Prior-session retrieval needs available, authorized history tools; the plugin does not install a history service. Earlier installation checks predate removal of named agent presets and do not validate the current bundle across all clients. During verification and repair, the agent runs project commands; inspect those commands and use test data first.
How to disable. Disable the client plugin and remove separately installed hooks if present. Saved .learnings and approved project rules remain for separate review.
FAQ
Must every task start with a long interview?
No. plan-interview is optional; a clear request does not require another goal-confirmation round.
Where are lessons stored?
In `.learnings/`, including `HEALS.md` for recorded fixes.
Who changes persistent rules?
The main agent proposes edits and the user approves them before they enter project instructions.
Does the bundle provide an MCP server?
No. It provides skills and native client integrations without a separate MCP server.
What if no useful history is found?
Continue with available evidence; the guidance does not require inventing a historical brief.
Discussion
No questions or reviews yet.
Related
PluginsA tool that builds tested skills from a real workflow and governs their lifecycle through your own marketplace
Self-learning
self-learning-skills
A meta-skill that saves a hard-won working procedure as a new skill or rule, so the next session starts already knowing it
A plugin of seven skills for Claude Code and Codex: prompts, subagents, skill authoring, editing, web search and a browser.
A self-improvement plugin for Claude Code, Codex and OpenCode: turns corrections into rules the agent follows in later sessions
A CLI that runs your agent on tasks and revises its skills, keeping a change only when held-out tasks confirm the gain
A collection of 14 portable skills for long agent runs: work planning, live app QA, search across past sessions and a phone call
README in short
Use skills individually or select them through `skill-pipeline`. The agent assigns independent reviews to general-purpose agents with read-only access.