andreaswasita/copilot-agents-dojo/durable-work
Claude Code and agent skill by andreaswasita/copilot-agents-dojo
- community
Checking that the SKILL.md file is still in the repository.
Picks the board over sub-agents for cross-turn work.
Install
npx vskill@latest install andreaswasita/copilot-agents-dojo/durable-workSKILL.md
Durable Work Skill
Routes work that must survive a single Copilot turn to tasks/board/ instead of the task tool. Sub-agents launched via task are killed if the parent is cancelled (rate-limit, user navigation, OS sleep) — anything you need across sessions belongs on the durable board. The complementary skill is skills/subagent-strategy, which handles in-turn delegation.
When to Use
- Work estimated > 1 turn (e.g. a multi-step deployment, a long build, an iterative review).
- Work that multiple agents or a human will hand off across sessions.
- Work whose state must be queryable later (
scripts/board.sh status). - Work whose loss would block the user (e.g. partial migrations, in-flight refactors).
- NOT for: one-shot research, fast file reads, parallel grep — those are sub-agent territory.
Prerequisites
tasks/board/exists (created byscripts/init.shor the dojo template).scripts/board.shavailable on PATH or asbash scripts/board.sh.- Familiarity with
.dojo/delegation.yaml→escalate_to_board_iftriggers.
How to Run
Quick Reference
| Step | Action | Tool | |
|---|---|---|---|
| Decide durable | Check .dojo/delegation.yaml triggers | view .dojo/delegation.yaml | |
| Create task | New board file from template | powershell → bash scripts/board.sh new "<title>" | |
| Status snapshot | Group tasks by status | powershell → bash scripts/board.sh list | |
| Single-line summary | Counts per status | powershell → bash scripts/board.sh status | |
| Update todo.md | Roll up board into the plan view | powershell → bash scripts/board.sh roll-up | |
| Edit task fields | Update frontmatter or checkboxes | edit tasks/board/<file>.md |
Procedure
Step 1: Decide Durable vs Ephemeral
Open .dojo/delegation.yaml → escalate_to_board_if. If ANY of those triggers fire, use the board. Otherwise, the work is in-turn and subagent-strategy applies. Do NOT default to the board — overuse turns it into a graveyard of stale tasks.
Step 2: Create a Board Entry
This copies tasks/board/000-template.md to the next ordinal (e.g. 001-migrate-legacy-auth-to-jwt.md), substitutes id/title/date, and prints the path.
Step 3: Fill Context and Plan
Open the new file. Write:
- Context — Why this task exists. Link the issue, the spec, the conversation.
- Plan — Concrete, verifiable checkboxes. NOT "improve auth" — write "replace cookie-based session check in
src/middleware/auth.jswith JWT verifier". - Verification — How we'll know it's done. Tests, diffs, smoke commands.
Step 4: Work and Update Status
Use the edit tool on the task file as work progresses. Update the status: field (pending → in_progress → done or blocked), bump updated:, and check off Plan items as they land. Append to Decisions as you make them — never rewrite history.
Step 5: Roll Up to todo.md
This regenerates tasks/todo.md from the board so anyone (human or agent) can see plan state at a glance. The board file remains the source of truth; todo.md is the rendered view. Run this before every session start and after every status flip.
Step 6: Close and Capture Lessons
When status: done, write the Lessons section. What would you tell a future agent picking up similar work? Cross-reference relevant tasks/lessons.md entries.
Pitfalls
- DO NOT use the board for in-turn research, parallel reads, or quick analysis — that's
subagent-strategy. The board has per-task overhead that swamps small work. - DO NOT edit
tasks/todo.mddirectly once the board is in use. Edits will be wiped on the nextroll-up. Edit the board file instead. - DO NOT leave stale
in_progresstasks — flip toblockedwith a reason if you stop. - DO NOT rename board files. The ordinal in the filename is the durable id; downstream tooling indexes it.
- DO NOT skip the Lessons section when closing a task — the board is also the dojo's institutional memory.
- DO NOT mix durable and ephemeral work in the same task file. One concern per file.
Verification
- [ ] Every task on the board has
id,title,status,owner,created,updatedfrontmatter. - [ ]
tasks/todo.mdmatches the board (runbash scripts/board.sh roll-upif not). - [ ] No
in_progresstask older than 2 days without aDecisionsupdate. - [ ] No edits to
tasks/todo.mdsince the lastroll-up. - [ ] Every
donetask has a non-emptyLessonssection.
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/durable-work/SKILL.md
Popularity
- Repository stars
- 52
- Repository forks
- 15
- 7-day trend score
- 20
Stars and forks are for the whole GitHub repository, not this skill alone. Refreshed 8 Sep 2026, 2 months ago.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/andreaswasita/copilot-agents-dojo/durable-work)