feat(beagle-analysis): add quick-plan skill (spec-free TDD planning)

[anderskev]

Summary

Adds quick-plan, a sibling to write-plan that produces the same bite-sized, TDD-driven implementation plan without requiring a brainstorm-beagle spec. It sources the contract from the live session instead: reconstructing intent from the conversation, fanning out domain-expert exploration subagents, and confirming with the user only where intent has real gaps.

Changes

Added

• plugins/beagle-analysis/skills/quick-plan/SKILL.md — the skill (218 lines).
• plugins/beagle-analysis/skills/quick-plan/references/fanout-brief.md — the explore-and-advise subagent dispatch template.

Changed

• CLAUDE.md / AGENTS.md — skill counts (131→132 total, beagle-analysis 13→14) and a Key Skills table row.
• CHANGELOG.md — [Unreleased] entry.

Motivation

write-plan hard-gates on a finalized spec at .beagle/concepts/<slug>/spec.md. That's the right discipline when a spec exists, but it blocks the common case where a feature has already been talked through in-session and the user just wants a plan. quick-plan fills that gap while preserving the planning rigor.

Design notes

A spec gives write-plan four things; quick-plan reconstructs each:

Spec provides	quick-plan reconstructs from
WHAT/WHY	Conversation → an Intent Brief written into the plan header
Key Decisions	Fanout domain-expert subagents recommending an approach grounded in real code
Reference Points	Same subagents returning file:line analogs
Human-reviewed gate	Adaptive gap check — silent when intent is clear, questions only for real holes

Deliberate calls:

• References write-plan's plan-template.md instead of forking it — single source of truth for the output format, so the two skills can't drift. quick-plan overrides only the header (an Intent block replaces the spec link) and inherits all TDD/self-review discipline.
• Tightens the spike + parallel-implementation gates — reconstructed intent carries more unverified claims than a vetted spec, so that's exactly where a no-spec plan is most likely to bake in a wrong assumption.
• Output goes to .beagle/plans/<slug>/plan.md (sibling to concepts/) to keep spec-derived and conversation-derived plans distinct.

Testing

This is a pure-markdown plugin marketplace (no build/test system). Verified manually:

• YAML frontmatter parses (Python yaml.safe_load).
• All 5 relative cross-references resolve (write-plan SKILL + template, brainstorm-beagle, subagent-prompt, local fanout-brief.md).
• SKILL.md is 218 lines (under the 500-line guideline).
• Documented skill counts match the on-disk count (14 in beagle-analysis).

Note: description-trigger optimization (skill-creator run_loop) was attempted but produced no usable signal in this environment — the harness exposes the description as a synthetic slash command (uniform 0/3, since natural-language prompts don't auto-invoke commands), and its improvement step requires ANTHROPIC_API_KEY, which isn't set. The hand-authored description includes explicit positive triggers, write-plan/brainstorm-beagle disambiguation, and negative cases.

Related Issues

None.

[coderabbitai]

Walkthrough

A new quick-plan skill is added to the beagle-analysis plugin. The skill reconstructs a plan intent from the current conversation when no .beagle/concepts/<slug>/spec.md exists, producing output at .beagle/plans/<slug>/plan.md. The specification defines a multi-step workflow: Intent Brief reconstruction, adaptive gap-check, parallel domain-expert fanout subagents, plan composition using the write-plan template (with a synthesized Intent block substitution), self-review checklist, user draft-approval gate, and subagent-prompt handoff generation. A companion fanout-brief.md reference document specifies the subagent dispatch protocol and report schema. AGENTS.md, CLAUDE.md, and CHANGELOG.md are updated to reflect the new skill count (132 total, 14 in beagle-analysis).

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly and concisely summarizes the main change: adding a new 'quick-plan' skill to beagle-analysis that works without requiring a spec.
Description check	✅ Passed	The description is detailed, well-structured, and directly relates to the changeset, explaining the motivation, design choices, and testing performed.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Caution

Some comments are outside the diff and can’t be posted inline due to platform limitations.

⚠️ Outside diff range comments (1)

CHANGELOG.md (1)

1-11: ⚠️ Potential issue | 🟠 Major

Update README.md skill counts to match AGENTS.md and CLAUDE.md.

AGENTS.md and CLAUDE.md have been updated to reflect 132 total skills and beagle-analysis with 14 skills (including the new quick-plan), but README.md still shows 131 active skills and beagle-analysis 13. Update the following in README.md:

• Line 9: Change "131 active skills" to "132 active skills"
• Line 64: Change "beagle-analysis | 13" to "beagle-analysis | 14"
• Line 66: Change Total | 131 to Total | 132

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@CHANGELOG.md` around lines 1 - 11, Update the README.md file to reflect the
new skill counts that were updated in AGENTS.md and CLAUDE.md. Change the total
active skills count from 131 to 132, update the beagle-analysis skill count from
13 to 14 in the skills table, and update the Total row in the same table from
131 to 132 to match the addition of the new quick-plan skill documented in the
CHANGELOG.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Outside diff comments:
In `@CHANGELOG.md`:
- Around line 1-11: Update the README.md file to reflect the new skill counts
that were updated in AGENTS.md and CLAUDE.md. Change the total active skills
count from 131 to 132, update the beagle-analysis skill count from 13 to 14 in
the skills table, and update the Total row in the same table from 131 to 132 to
match the addition of the new quick-plan skill documented in the CHANGELOG.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 058cd5d7-4f84-4b50-af54-e9e8fe505e08

📥 Commits

Reviewing files that changed from the base of the PR and between 4b57133 and 22c6eeb.

📒 Files selected for processing (5)

• AGENTS.md
• CHANGELOG.md
• CLAUDE.md
• plugins/beagle-analysis/skills/quick-plan/SKILL.md
• plugins/beagle-analysis/skills/quick-plan/references/fanout-brief.md