Done

Archive completed work to Done.md and browse past accomplishments.

Config Preamble

Before any operation, locate the PlanKit folder and read its config:

1. Check if PlanKit/ exists in the project root. If not, check plan-kit/.
2. Read .config.json from the PlanKit folder to get the naming convention.

Determine file path:

• Done: PlanKit/Done.md or plan-kit/done.md

If Done.md doesn't exist, create it with just # Done as content.

Archive Scenarios

Other PlanKit skills trigger these scenarios by loading this skill and specifying which one applies. When loaded, find the matching scenario and follow its instructions.

Scenario 1: Step Completed (or Dropped)

Triggered by the detail-steps skill after a step file is deleted.

Input from calling skill: step name, step letter, feature name, feature number, file path, status (✅ completed or 🛑 dropped), reason (for dropped only), PR numbers (if known).

1. Get the deletion commit SHA: git log -1 --format=%h -- "path/to/deleted/file"
   • This finds the commit that last touched the file (the deletion commit)
2. Read Done.md
3. Find or create the version section:
   • Look for ## vX.Y matching the feature's roadmap version
   • If it doesn't exist, create it: ## vX.Y — Subtitle (YYYY-MM-DD →)
   • Start date = today if this is the first entry for this version
   • End date left blank (filled when version is released)
4. Find or create the feature subsection:
   • Look for ### ... NNN-FeatureName under the version section
   • If it doesn't exist, create it with the feature's one-line description from the Roadmap
   • Don't set the feature-level status marker yet (it's set when the feature is fully complete)
5. Add the step line:
   • Format: - ✅ X-StepName · plan: \SHA`` (for completed)
   • Format: - 🛑 X-StepName · plan: \SHA` — reason` (for dropped)
   • Append · PR #N if PR numbers were provided
6. Write Done.md

Scenario 2: Feature Completed

Triggered by the detail-steps skill after all steps are done and the feature folder is deleted.

1. Read Done.md
2. Find the feature subsection (should already exist from Scenario 1 step entries)
3. Set the feature status marker to ✅: update ### line to ### ✅ NNN-FeatureName
4. Verify all steps are listed — if any are missing (completed before Done.md existed), add them now
5. Write Done.md

Scenario 3: Feature Skipped or Dropped

Triggered by the plan-roadmap skill when a feature is removed from the roadmap without completion.

Input: feature name, feature number (if it had one), version, status (⏭️ skipped or 🛑 dropped), reason.

1. Read Done.md
2. Find or create the version section (same as Scenario 1, step 3)
3. Add the feature entry:
   • ### ⏭️ NNN-FeatureName (skipped — never started) or ### 🛑 NNN-FeatureName (dropped — started but abandoned)
   • If the feature never got a number (removed during roadmap triage before steps), use just the feature name without a number
   • Next line: the reason, as plain text
4. Archive any existing steps — if the feature had step files, archive each one with its actual status
5. Write Done.md

Scenario 4: Version Released

Triggered by the plan-roadmap or dashboard skill when a version ships.

Input: version number, subtitle, release date, list of features with their final statuses.

1. Read Done.md
2. Find the version section (should already exist from earlier step/feature entries)
3. Fill in the end date: update (YYYY-MM-DD →) to (YYYY-MM-DD → YYYY-MM-DD)
4. Ensure all features have final status markers (✅, ⏭️, or 🛑)
5. Add any features that aren't listed yet (features without steps that were completed directly)
6. Write Done.md

Browse Mode

When the user wants to review completed work (triggered by keywords like "what did we finish", "show history", "browse done"):

1. Read Done.md
2. Present a summary by version: version name, date range, feature count, status distribution
3. Offer to show details for specific versions or features
4. For recovery: when user wants to see a deleted step file, use the plan: SHA:
   bashCopy

   git show <sha>~1:PlanKit/Features/NNN-FeatureName/X-StepName.md
   

   (Use ~1 to get the parent commit, which still has the file content.)

Important Rules

• Done.md is append-mostly — entries are added as work completes, rarely modified after creation
• Version sections ordered newest-first — newest version at the top, oldest at the bottom
• Git SHAs enable full recovery — every deleted step file can be reconstructed from its plan: SHA
• PR references are optional — include when the user mentions them, skip otherwise. Can be added retroactively.
• Reason is mandatory for ⏭️ and 🛑 — always explain why something was skipped or dropped
• One-line descriptions — feature descriptions are a single sentence from the Roadmap, not the full detail
• Cross-references use NNN/X format — e.g., 005/A for feature 005, step A

Format Reference

For detailed file format specification, see DONE_FORMAT.md.