aide-handoff
Skill4 starsChanged 18 days agoMITView on GitHub
Record current pipeline state to `.aide/session.aide` — the project-wide pipeline-position log that lets the next session (yours or another agent's) resume from a known point. The file always exists, starting as an empty placeholder; every handoff is an UPDATE that writes whatever the agent has, with or without a task. `session.aide` is always kept up to date as the agent works. This skill is the orchestration layer: it gathers content from the invoking agent's conversation context and delegates the actual write to the `aide-spec-writer` agent in `session.aide` mode. Invoke this skill when the user types `/aide:handoff`, asks to "save the session" or "record where we are", and whenever anything changes what `session.aide` says — a stage starts or completes, a decision is settled, an invariant is added, a task moves, a release ships, or work outside the pipeline changes the state or the next step. Never end a turn or a session with `session.aide` behind. `session.aide` is never deleted — when a feature closes, UPDATE it for the next task.
In this repository6 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- aide-brain-authorSkill
- aideSkill
- brainSkill
- kanbanSkill
- study-playbookSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
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 40
when a task was created, moved, finished, or newly tied to the work, the matching ## Kanban edit
≈ “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
/aide:handoff — Session Handoff
Record current pipeline state to .aide/session.aide. This skill owns the orchestration; the aide-spec-writer agent owns the actual file write (it transcribes orchestrator-supplied state into the canonical format). Nothing deletes session.aide; it is long-lived and rewritten as features close. See session.aide spec for the file format.
Hard constraints
- You do NOT write
.aide/session.aideyourself. The actual file write belongs toaide-spec-writer. This skill describes the orchestration that surrounds the delegation. - You do NOT invent content. The invoking agent (orchestrator or other) holds the conversation context — current stage, settled decisions, where work paused, anti-regression invariants. That context is the content source. Write what it supplies; leave the rest empty — never fabricate state, and never hold the write back for content you don't have.
- You do NOT delete
.aide/session.aide. Nothing does. This skill is exclusively for create/update.
Checklist
- Make sure the file exists. It always should —
aide_initand the/aideboot create the empty placeholder withaide_scaffold type=session. If.aide/session.aideis missing anyway, callaide_scaffold({ directory: ".", type: "session" })first — the placeholder's single source; it never overwrites. Then continue. Gather content from conversation context. Every handoff is an UPDATE. Supply whatever you have, section by section:
- the change description (what happened)
- section-targeted edits, e.g. "fill the intent with: ...", "append decision #4: ...", "rewrite
## Where this cycle stoppedto: ..." - numbered references to preserve (decision #N stays stable across edits)
- when a task was created, moved, finished, or newly tied to the work, the matching
## Kanbanedit
The first handoff after the placeholder fills in whatever is known: the one-line intent, the state summary, where the cycle stopped and the next steps, what is done, decisions, invariants, open questions. A missing piece is not a reason to wait. Write what you have now; empty sections stay empty until there is something true to put in them.
- Keep it the full session record.
session.aideholds what was done, every decision, what is validated and what comes next, in full. A fresh agent reading only this file must know where everything stands. Stashing into the task through thekanbanskill still happens alongside, but never in place of the detail here. - Link tasks when there are tasks. Every task tied to the work goes into
## Kanbanby slug, column and what it covers. The slug is the stable link — the kanban skill finds the task by slug in whatever column it sits. Work with no task is still recorded: never hold a handoff back until a task exists. If it is trackable work, ask the user whether to create a task for it, and link it once it exists. - Delegate to
aide-spec-writervia the Agent tool (subagent_type: aide-spec-writer). The delegation prompt MUST:- Name the operation type explicitly —
session.aide UPDATE. The spec-writer's default operation is.aidefrontmatter; naming the operation is mandatory for correct dispatch. - Quote all supplied content verbatim — the spec-writer transcribes; it never invents.
- Name the operation type explicitly —
- Receive the verdict — UPDATED. A REFUSED only means the prompt was ambiguous about which section an edit belongs to; clarify and re-delegate.
- Confirm completion and continue with whatever pipeline work was in flight before the handoff was invoked.
What this skill does NOT do
- It does not interview the user. The invoking agent holds the conversation context; this skill only formats it into a delegation prompt for
aide-spec-writer. - It does not apply the Brevity Contract.
session.aideis operational state, not durable intent — no caps on description, no word counts, no forbidden-content rules from the.aidebody. The Brevity Contract is.aide-only. - It does not delete
session.aide. Nothing does — at feature close, UPDATE it for the next task. - It does not bypass
aide-spec-writer. The skill is the orchestration layer; the agent owns the file write. CallingWrite/Editon.aide/session.aidedirectly from this skill violates the delegation pattern.
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.

