agents-course plugin
Tools from the agents course: audit your context, write done as a script, bootstrap your eval suite, and a loop critic agent.
How to install the agents-course plugin
Run both commands inside Claude Code. The first adds the marketplace once; the second installs the plugin.
/plugin marketplace add undefined-ui/second-brain-os
/plugin install agents-course@second-brain-osWhat's inside agents-course
6 parts read from the plugin's own files.
Agents 1
Call as @agent-agents-course:name- @agent-agents-course:loop-critic
>-
Skills 5
Claude loads it when the task matches- context-audit
Audit an agent's context layout against the four places: system prompt, tools, history, tail. Use when the user asks to audit, review or fix their agent's context, prompt caching, token spend per turn, or CLAUDE.md / AGENTS.md layout. Read-only: reports findings and a fix list, changes nothing. Do NOT use for writing evals (evals-bootstrap) or for loop design (goal-test).
- evals-bootstrap
Scaffold a first eval suite for an agent: mine real failures into cases, write behavioural checks over traces, and generate the runner. Use when the user wants evals, regression tests for an agent, a golden set, or asks how to know a prompt/model change did not break things. Do NOT use for a single task's done-check (goal-test) or for auditing context (context-audit).
- gate-check
Find the decisions in a pipeline that do not need the expensive model and propose the gate for each: a rule, a classic classifier, or a small model, with fail-closed routing. Use when the user asks to cut model costs or latency, says the big model handles everything, or wants a triage / routing / filter layer in front of an agent. Analysis plus an optional baseline scaffold. Do NOT use for auditing context layout (context-audit) or for building eval suites (evals-bootstrap).
- goal-test
Turn a vague task into a testable definition of done and generate an executable goal-test script for it, optionally with a bounded retry loop around a headless agent. Use when the user wants to run an agent in a loop, asks how to know when an agent task is finished, or says a goal like "improve X" needs to become checkable. Do NOT use for building eval suites over many cases — that is evals-bootstrap.
- harness-audit
Audit an agent's harness against the four rings: containment, guides, sensors, permissions. Use when the user asks whether their agent setup is safe, wants a review of sandboxing / permissions / hooks / AGENTS.md guardrails, or asks what a bad session could break. Read-only: reports the blast radius and a hardening list, changes nothing. Do NOT use for context layout (context-audit) or cost routing (gate-check).
Using agents-course in Claude Code
Ask for an agent by name:
@agent-agents-course:loop-criticSkills need no command. Claude reads them when your request matches what they describe.
Update, turn off or remove
- Update the plugin (shell)
- claude plugin update agents-course@second-brain-os
- Refresh the marketplace
- /plugin marketplace update second-brain-os
- Turn it off for now
- /plugin disable agents-course@second-brain-os
- Remove it
- /plugin uninstall agents-course@second-brain-os
Skills in agents-course
- context-auditundefined-ui/second-brain-os
Audit an agent's context layout against the four places: system prompt, tools, history, tail. Use when the user asks to audit, review or fix their agent's context, prompt caching, token spend per turn, or CLAUDE.md / AGENTS.md layout. Read-only: reports findings and a fix list, changes nothing. Do NOT use for writing evals (evals-bootstrap) or for loop design (goal-test).
1.6Kstars - evals-bootstrapundefined-ui/second-brain-os
Scaffold a first eval suite for an agent: mine real failures into cases, write behavioural checks over traces, and generate the runner. Use when the user wants evals, regression tests for an agent, a golden set, or asks how to know a prompt/model change did not break things. Do NOT use for a single task's done-check (goal-test) or for auditing context (context-audit).
1.6Kstars - gate-checkundefined-ui/second-brain-os
Find the decisions in a pipeline that do not need the expensive model and propose the gate for each: a rule, a classic classifier, or a small model, with fail-closed routing. Use when the user asks to cut model costs or latency, says the big model handles everything, or wants a triage / routing / filter layer in front of an agent. Analysis plus an optional baseline scaffold. Do NOT use for auditing context layout (context-audit) or for building eval suites (evals-bootstrap).
1.6Kstars - goal-testundefined-ui/second-brain-os
Turn a vague task into a testable definition of done and generate an executable goal-test script for it, optionally with a bounded retry loop around a headless agent. Use when the user wants to run an agent in a loop, asks how to know when an agent task is finished, or says a goal like "improve X" needs to become checkable. Do NOT use for building eval suites over many cases — that is evals-bootstrap.
1.6Kstars - harness-auditundefined-ui/second-brain-os
Audit an agent's harness against the four rings: containment, guides, sensors, permissions. Use when the user asks whether their agent setup is safe, wants a review of sandboxing / permissions / hooks / AGENTS.md guardrails, or asks what a bad session could break. Read-only: reports the blast radius and a hardening list, changes nothing. Do NOT use for context layout (context-audit) or cost routing (gate-check).
1.6Kstars
agents-course plugin questions
How do I install the agents-course plugin?
In Claude Code run /plugin marketplace add undefined-ui/second-brain-os, then /plugin install agents-course@second-brain-os.
What does the agents-course plugin include?
agents-course bundles 1 agent, 5 skills.
Which agents come with agents-course?
loop-critic. Call one with @agent-agents-course:loop-critic.
How do I update or remove the agents-course plugin?
Update it with claude plugin update agents-course@second-brain-os in your shell, or from the Installed tab of /plugin. Remove it with /plugin uninstall agents-course@second-brain-os.
What license is the agents-course plugin under?
MIT, as stated in the plugin's manifest.