agentsclimarketplace

Agent team discuss

Skill mike-diff/ai-coding-configs/.pi/skills/agent-team-discuss

Multi-agent harness configurations: agent teams, safety hooks, skills and context engineering patterns

Install
npx -y skills add mike-diff/ai-coding-configs --skill agent-team-discuss

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Run the Agent Team discuss workflow in pi for this repo. Use when exploring a feature idea before writing a spec or implementation plan, validating assumptions, or preparing an ADLC handoff for /skill:agent-team-spec or /skill:agent-team-dev.

SKILL.md

2.7 KB, as published. Nobody here has run it

Agent Team Discuss

Use this skill to run the Agent Team /discuss workflow from pi while maintaining repo parity.

Source of truth

The product workflow source of truth is:

  • .claude/skills/discuss/SKILL.md
  • .claude/skills/discuss/references/phases.md
  • .cursor/skills/discuss/SKILL.md
  • .cursor/skills/discuss/references/phases.md

Read those files when details matter. This pi skill is a maintainer/operator wrapper, not a fourth independent workflow implementation.

Goal

Answer: Should we build this, what are we assuming, and what should flow into /skill:agent-team-spec or /skill:agent-team-dev?

Process

  1. Parse the idea and detect mode:
    • Fresh idea
    • Revisit existing implementation
    • Reference-driven discussion using files, docs, URLs, or prior context
  2. Ask at most 3 clarifying questions at a time.
  3. Use repo search and file reads to ground the conversation in existing patterns.
  4. Do the research inline, sequentially (pi is single-agent — there is no team/subagent spawn):
    • Scout the codebase for implementation patterns
    • Research external libraries or current platform behavior (web/docs)
    • Stress-test high-risk plans for blind spots For heavy parallel research, the multi-agent COUNCIL flow lives in the source-of-truth .claude/skills/discuss/references/phases.md / .cursor — run it there.
  5. Blind-spot check is MANDATORY before recommending build — never skip it, and do not gate it on "high-risk only" (per the source-of-truth phases.md).
  6. Validate the plan before recommending build.
  7. Do not implement.
  8. End with a compact validated plan and an <adlc-handoff> block.

Required final handoff

<adlc-handoff>
problem: [one-sentence problem]
target_user: [primary user or actor]
hypothesis: [why this approach should work]
success_metrics:
  - [measurable success signal]
core_workflow_break: [what is broken or missing today]
assumptions:
  - [assumption to validate]
risks:
  - [risk to mitigate]
human_decisions_required:
  - [decision user must make before build, or None]
recommended_next: /skill:agent-team-spec "[feature]" or /skill:agent-team-dev "[small validated change]"
</adlc-handoff>

Guardrails

  • Keep the public workflow lightweight. Do not introduce new ADLC commands.
  • If a spec is needed, recommend .context/specs/spec-[feature-name].md.
  • If you notice drift between .claude/, plugin and .cursor/, flag it before proceeding.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.