linear-task-lifecycle
Skill1 starsChanged 2 months agoMITView on GitHub
Manage a Linear issue through implementation, verification, interruption, and completion. Use when a coding task references a Linear issue identifier or URL, or when the user asks an agent to implement, resume, review, or update tracked Linear work. Do not mutate Linear for read-only questions unless explicitly requested.
In this repository15 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- codebase-inventorySkill
- code-commentsSkill
- codex-goal-writerSkill
- generate-specSkill
- git-commitSkill
- glab-mr-reviewSkill
- glab-mrSkill
- handoff-auditSkill
- jira-apiSkill
- reject-and-redoSkill
- setup-serenaSkill
- setup-skill-syncSkill
- subagent-goSkill
- wtSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
17 rules. About 0 are common, 16 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈16
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 8
Treat Linear as the source of truth for task scope, status, dependencies, and delivery discussion.
- Distinctive: line 12
When the request identifies a Linear issue, fetch the issue before acting.
- Distinctive: line 14
If the request is only consultation, planning, or a read-only review, do not change the issue unless the user explicitly asks.
- Distinctive: line 21
If a blocking relation is unresolved, do not begin implementation.
- Distinctive: line 23
Add one concise start or resume comment containing the branch or worktree when known and the intended validation.
- Distinctive: line 29
When the issue belongs to a Linear Project, treat Project status as an aggregate lifecycle rather than mirroring the current issue:
- Distinctive: line 31
When implementation of the first active issue actually begins, move an unstarted Project to its started status when the workspace exposes …
- Distinctive: line 32
Do not complete, cancel, or pause a Project merely because one issue changes state.
- Distinctive: line 33
Before completing a Project, fetch all Project issues and confirm that every required, non-canceled issue is completed and that any …
- Distinctive: line 34
When the Project is genuinely complete, publish one concise Project update with delivery and verification evidence, then move it to the …
- Distinctive: line 35
If Project status cannot be changed through the available Linear tools, report that limitation.
- Distinctive: line 37
Use the workspace's actual Project statuses.
- Distinctive: line 41
Use comments for durable handoff information, not a transcript.
- Distinctive: line 51
Do not mark an issue blocked merely because the work is difficult, incomplete, or awaiting the agent's own next step.
- Distinctive: line 55
Do not move the issue to a completed status until all of the following are true:
- Distinctive: line 61
no known required work remains.
- Distinctive: line 65
If implementation is delivered but required verification is still outstanding, keep the issue in a started state and state the remaining …
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Linear Task Lifecycle
Treat Linear as the source of truth for task scope, status, dependencies, and delivery discussion. This skill complements implementation, review, commit, and merge-request skills; it does not replace them.
Resolve the task
When the request identifies a Linear issue, fetch the issue before acting. Read its description, project, status, relations, and relevant comments. Use the issue's team and its actual workflow statuses rather than assuming status names.
If the request is only consultation, planning, or a read-only review, do not change the issue unless the user explicitly asks. If no issue identifier or URL is supplied, do not guess, search for a likely issue, or create one merely to track the work.
Start or resume implementation
An explicit request to implement or resume a Linear issue authorizes normal lifecycle updates to that issue:
- Confirm that the issue is not completed, canceled, or blocked by an unresolved dependency.
- If a blocking relation is unresolved, do not begin implementation. Report the blocker and leave an explanatory comment only when it adds information not already present.
- Move an unstarted issue to the team's started status immediately before implementation begins. Preserve an existing started status when resuming.
- Add one concise start or resume comment containing the branch or worktree when known and the intended validation. Do not post routine progress narration.
The issue defines scope. If repository evidence conflicts with it or required information is missing, pause and ask a targeted question instead of silently changing scope. Record the resulting decision in Linear when it affects future work.
Keep the project state consistent
When the issue belongs to a Linear Project, treat Project status as an aggregate lifecycle rather than mirroring the current issue:
- When implementation of the first active issue actually begins, move an unstarted Project to its started status when the workspace exposes an appropriate transition.
- Do not complete, cancel, or pause a Project merely because one issue changes state.
- Before completing a Project, fetch all Project issues and confirm that every required, non-canceled issue is completed and that any Project-level acceptance criteria are satisfied.
- When the Project is genuinely complete, publish one concise Project update with delivery and verification evidence, then move it to the completed status.
- If Project status cannot be changed through the available Linear tools, report that limitation; do not claim it was synchronized.
Use the workspace's actual Project statuses. Do not invent a status or substitute an Issue status for a Project status.
Record meaningful state
Use comments for durable handoff information, not a transcript. Record only material events such as:
- a user-approved scope or acceptance-criteria change;
- a newly discovered external blocker;
- validation evidence needed to judge completion;
- a delivery reference such as commit, branch, or merge request;
- recovery context before an intentional stop or context reset.
A recovery comment should state what is complete, what remains, the current branch or worktree when known, validation already run, and the next concrete action.
Do not mark an issue blocked merely because the work is difficult, incomplete, or awaiting the agent's own next step. Use the team's blocked state only for a genuine external dependency or missing user decision, and explain the condition in a comment.
Complete the issue
Do not move the issue to a completed status until all of the following are true:
- the issue's requested scope and acceptance criteria are satisfied;
- required source verification has passed;
- required runtime or human verification has actually been performed when the issue requires it;
- the requested commit or delivery operation is complete;
- no known required work remains.
On completion, add one delivery comment summarizing the implementation, commit or merge-request reference when available, and exact verification results. Then move the issue to the team's completed status.
If implementation is delivered but required verification is still outstanding, keep the issue in a started state and state the remaining gate. Never report a Linear transition that the MCP did not confirm.
Typical invocation
- Claude Code and Antigravity CLI (
agy):/implement BOL-5 - Codex:
$implement BOL-5
The implementation skill handles the code workflow; this skill fetches BOL-5, maintains its lifecycle state, and records durable evidence.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

