agentleFS
ElguajoElguajo/Progressive-Context-Kit/.agents/skills/session-handoff/SKILL.md

session-handoff

Skill0 starsChanged 42 days agoMITView on GitHub

End of a meaningful implementation/review session or phase transition.

139 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use ascii-diagram, research-fleet and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • SethGammonSkill

    ascii-diagram

    Generate perfectly aligned ASCII diagrams — architecture, flow, sequence, box-and-arrow. Uses a programmatic character-grid approach so alignment is guaranteed by math, not token prediction. Includes post-render verification.

    Also in 4 projects that use this

  • SethGammonSkill

    research-fleet

    Parallel research using Fleet wave mechanics. Spawns multiple scout agents, each investigating a different angle of the same question. Findings are compressed between waves. Produces a unified research brief from multiple independent perspectives.

    Also in 3 projects that use this

  • vicoa-aiSkill

    live-preview

    Start a local app in the current user project, expose it through a tunnel, and return preview details in structured Markdown. Use when the user asks for a preview URL, public URL, shareable localhost, mobile/web preview, or tunnel setup.

    Also in 4 projects that use this

  • botlearn-aiSkill

    doc-gen

    Also in 4 projects that use this

  • flyingobsidianSkill

    terminology

    Read a codebase, find terminology, create/update definitions into a markdown document.

    Also in 3 projects that use this

  • Dizzy-TenSkill

    bug-diagnosis

    Diagnose bugs and product issues without modifying code. Use when the user explicitly invokes $bug-diagnosis or asks for diagnosis, root-cause analysis, issue classification, or fix strategy for code bugs, failing tests, configuration/dependency problems, deployment issues, performance regressions, data anomalies, or frontend/user-experience problems described through screenshots, annotations, logs, errors, or natural-language reports. Prioritize analysis and recommendations only; do not edit files.

    Also in 3 projects that use this

  • +1 more from SethGammon

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 2 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 117 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read current acceptance evidence and docs/system/HANDOFF_PROTOCOL.md.

    line 10
  2. Very specific:

    Use Single-Focus Continuation for NEXT_SESSION: one handoff prompt = one unresolved execution target.

    line 34
  3. Distinctive:

    Keep the check narrow: do not replay completed work or mine chat/transcript history unless a specific contradiction requires cold …

    line 47
  4. Distinctive:

    Do not repeat the diff line by line, claim success beyond observed evidence, or ask for post-hoc approval after a clean finish unless …

    line 51

“Common” is measured: kept by at least 6 in 10 of the 117 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Session Handoff

Read current acceptance evidence and docs/system/HANDOFF_PROTOCOL.md. Classify the session as IN PROGRESS, PHASE COMPLETE, or PROJECT COMPLETE. On PHASE COMPLETE, update canonical owners first, write the durable phase completion report, then write the completed phase's compact ## Completion Record pointing to that report before updating Roadmap markers. Retain Phase-owned task requests/prompts and observed outcome/evidence; reference their IDs in the report. Update NEXT_SESSION only when hot navigation/state changes, using an Active task prompt pointer instead of an editable executable prompt.

The phase completion report may preserve evidence-bounded technical detail useful to humans or later investigation. The Completion Record must remain small enough for normal progressive warm-up and must not duplicate the report. Legacy completed phases that predate separate completion reports remain valid and do not require automatic migration.

For a safe mid-work pause, remain IN PROGRESS: finish the current atomic unit or return it to a known recoverable state, start no queued unit, and record only actual disk/project state plus observed verification. State whether it is RUNNABLE / GREEN, KNOWN BROKEN / RECOVERABLE (with the exact issue and first recovery action), or BLOCKED. Do not fabricate completion artifacts, create another ledger, or require a WIP Git commit.

For a non-trivial task inside an active phase, preserve its request/prompt, freshness checks, observed result/evidence and relevant decisions/issues in its Phase-owned archive record. Keep the same ID for unchanged instructions; changed instructions require a replacement revision linked to the retained old record. Do not create one completion file per routine task.

Use Single-Focus Continuation for NEXT_SESSION: one handoff prompt = one unresolved execution target. If the current task, acceptance criterion, manual verification, blocker, or other gate remains open, both Next action and the referenced Phase prompt must focus only on closing that target. Do not name or preload the next queued task/phase with wording such as "then continue" or "after that start". Multiple substeps are allowed only when they all complete the same target and share one acceptance boundary. Select later work only after the current target is genuinely complete and its evidence/state has been persisted.

On a cold-start continuation, treat NEXT_SESSION as a pointer rather than authority. Reconcile its target with the Roadmap/current Phase, live worktree/repository state, and decisive evidence before acting. Apply the freshness check in docs/system/CONTEXT_PROTOCOL.md, even for a record previously marked CURRENT. Preserve stale prompts and revise navigation/instructions when those sources contradict them; migrate legacy inline prompts only when continuing them. Keep the check narrow: do not replay completed work or mine chat/transcript history unless a specific contradiction requires cold historical evidence.

Do not repeat the diff line by line, claim success beyond observed evidence, or ask for post-hoc approval after a clean finish unless another risky step remains. Include the short launch instruction pointing to the saved Phase task ID and requiring a freshness check when the project continues; do not repeat the task prompt body. Never end a turn with only "waiting for confirmation/next action is X" — persist state and provide the continuation record pointer before yielding to the user.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.