agentleFS
Jamie-BitFlightJamie-BitFlight/claude_skills/.agents/skills/rebase/SKILL.md

rebase

Skill67 starsChanged 52 days agoMITView on GitHub

Start a local Git rebase when the user explicitly requests replay of a named source ref, or of the current branch, onto a named target, including a rebase-then-push request, which carries exact-lease publication to the branch's own remote destination, or continue or abort an active rebase. Use only when local history replay will start or is active; do not use for merge-based branch updates, repository merge settings, pull-request or merge-request merging, or push-only requests.

27 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use dependency-watch, perform-task 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.
  • telegramdesktopSkill

    dependency-watch

    Audit Telegram Desktop dependencies on freshly fetched origin/dev for releases and security fixes, including upstream lag and backport candidates in patched forks. Use for the scheduled dependency monitor or an explicitly requested dependency report.

    Also in 5 projects that use this

  • telegramdesktopSkill

    perform-task

    Resolve, start or resume, implement, review, test, and publish exactly one existing ai-tdesktop task by short slug or full dated id, including rare blocked retries and split-required results. Use when the user invokes $perform-task or /perform-task with a known task name, or when the continue scheduler delegates one selected task. Runs standard review lenses with fast applicability bailouts and selects task-specific domain and evidence instruments without selecting additional work.

    Also in 5 projects that use this

  • JerovinoSkill

    process-inbox

    Processes tasks.md inbox items by triaging each one to its proper location in the vault

    Also in 5 projects that use this

  • Trejon-888Skill

    continue

    Load project context, alignment audit, status check, suggest next action. Auto-invoke at session start, after context switches, or when user says 'continue', 'pick up where we left off', 'resume', 'what's next', 'status check', 'what now'.

    Also in 6 projects that use this

  • lepinkainenSkill

    code-health

    Analyze codebase health - large files, test coverage gaps, duplicate code, dead/legacy code, and documentation issues. Use when asked to scan, audit, or assess code quality, find refactoring targets, or identify technical debt.

    Also in 3 projects that use this

  • dekozaSkill

    fix-pr

    Repair an existing pull request's review findings and target-branch conflicts, verify the result, and update the same open PR.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 27 edited versions we read keep these.

  • 0

    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:

    Run repository-required and changed-interaction checks on a clean detached worktree or equivalent isolated checkout of R, separately from …

    line 11
  2. Very specific:

    Every result change, including remote reconciliation, invalidates earlier result validation.

    line 18

“Common” is measured: kept by at least 6 in 10 of the 27 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

Keywords: rebase, git rebase, history replay, rebase conflict, continue rebase, abort rebase, git worktree, rewritten history, authorized force-with-lease publication

Result-bound verification

For start/continue completion, bind validation to the immutable result commit R. Run repository-required and changed-interaction checks on a clean detached worktree or equivalent isolated checkout of R, separately from restored user changes. Preserve unrelated work; a passing dirty worktree does not validate its commit. Record the checked OID and environment; verify checks left tracked source unchanged. Source-mutating checks invalidate that evidence. Commit a required correction within the existing task authority, then rebind R and rerun its checks. When that commit is blocked or unauthorized, pause rather than reporting the correction published. Every result change, including remote reconciliation, invalidates earlier result validation. Reobserve the named result ref before completion; it must still name R. A decision or stopped-state path cannot become a success merely because its refs or ancestry still satisfy the goal.

Publication authority

A start request to rebase a branch authorizes publishing R to that branch's own remote destination, using --force-with-lease bound to an observed destination OID. Local-only wording in the request withdraws it. Continue publishes only under authority bound at start; abort never publishes. Any other destination, any lease not bound to an observed OID, and any other force-push needs explicit authority.

Other parties update remote feature branches. Fetch the destination before replay and again before pushing; integrate what it gained, then continue (publication).

Workflow

flowchart TD
    Start([Agent receives explicit start, continue, or abort request]) --> Kind{Request?}
    Kind -->|Start| Bind[Agent: from request/task context bind exact refs/OIDs, requested observable history/ref relation, destinations, authority, worker facts]; Kind -->|Continue or abort| Active["`Agent: read [active recovery](./references/active-rebase-recovery.md); bind metadata-owning worktree/Git dir`"]
    Bind --> BindResult{All dependent facts observed or explicitly bound?}
    BindResult -->|Yes| Satisfied{No active rebase; S/T unchanged; ancestry-only; T ancestor of S?}; BindResult -->|Missing or ambiguous| Decision; BindResult -->|Failure or unobservable| Recover
    Satisfied -->|Yes; destinations equal required result; read-only| NoChange[Agent: record no-change evidence]; Satisfied -->|No or transformation goal| Locate
    Satisfied -->|Yes; only publication differs| Locate; Satisfied -->|Failure or unobservable| Recover
    Active --> ActiveResult{Inspection result?}; ActiveResult -->|No metadata in bound worktree and Git dir| NoActive[Agent: record absent metadata and no mutation]; ActiveResult -->|Owner and required contract rebound| Safe
    ActiveResult -->|Required contract fact missing or inconsistent| Decision; ActiveResult -->|Failure or unobservable| Recover
    Locate[Agent: find exact source attachment; record worktree/Git-dir ownership] --> Location{Bound source location?}
    Location -->|Source branch attached here| Safe; Location -->|Source branch unattached| BranchWT[Agent: create start-only branch worktree]
    Location -->|Source attached elsewhere| BindWT[Agent: bind that exact worktree/Git dir]; Location -->|Non-branch OID; result bound| DetachedWT[Agent: create start-only detached worktree]
    Location -->|Failure or unobservable| Recover; BindWT --> Safe
    BranchWT --> WTResult{Creation result?}; DetachedWT --> WTResult
    WTResult -->|HEAD OID equals source; branch attachment matches when required| Safe; WTResult -->|Failure or unobservable| Recover
    Safe[Agent: observe worktree writers; request pause when supported] --> Owner{Worker checkpoint observation?}
    Owner -->|No writer, acknowledged pause, or worker ended| Operation{Operation?}; Owner -->|Writer active| Wait[Agent: wait once for pause, end, or command result]
    Owner -->|Unknown or unreachable| Decision
    Wait --> WaitResult{Observed wait outcome?}; WaitResult -->|Checkpoint met or worker ended/failed without writer| Operation; WaitResult -->|Still running and cannot pause/end| Decision
    WaitResult -->|Failure or unobservable| Recover
    Operation -->|Start| Prepare[Agent: observe status; checkpoint only task-authorized changes after worker checkpoint]; Operation -->|Continue or abort| ActiveGuard[Agent: reobserve same owner metadata and saved-entry identity]
    ActiveGuard --> ActiveGuardResult{Owning worktree/Git-dir metadata unchanged?}
    ActiveGuardResult -->|Continue| Stop{Current active-rebase state?}; ActiveGuardResult -->|Abort| Abort["`Agent: read [active recovery](./references/active-rebase-recovery.md); abort in bound owner and restore pre-state`"]
    ActiveGuardResult -->|No, failure, or unobservable| Recover
    Prepare --> PrepResult{Status empty after checkpoint containing every observed change?}
    PrepResult -->|Yes| Refs; PrepResult -->|No; checkpoint commit blocked| Save["`Agent: read [named stash](./references/named-stash.md); save and bind exact entry`"]
    PrepResult -->|Failure or unobservable| Recover
    Save --> SaveResult{Exact entry verified?}; SaveResult -->|Yes| Refs; SaveResult -->|No or unobservable| Recover
    Refs["`Agent: reobserve exact source and target; when a publication destination is bound, fetch it per [publication](./references/publication.md) remote sync`"] --> RefResult{Ref result?}
    RefResult -->|Bound OIDs unchanged| Next; RefResult -->|Same named refs or bound destination moved| Account[Agent: account for movement; integrate compatible destination commits into the source before replay]
    RefResult -->|Failure or unobservable| Recover
    Account --> AccountResult{Every moved commit has one evidence-backed disposition?}
    AccountResult -->|All classified; no supported conflict| Next; AccountResult -->|Unclassified or conflicting| Decision
    AccountResult -->|Failure or unobservable| Recover
    Next{Reobserved start path?} -->|Replay required| Shape; Next -->|No-replay publication; bound goal relation true| Current[Agent: bind current result OID R]
    Next -->|Failure or unobservable| Recover
    Current --> CurrentResult{Named result ref resolves to R?}; CurrentResult -->|Yes| Saved; CurrentResult -->|No or unobservable| Recover
    Shape{Merge in fresh replay set?} -->|Yes| History["`Agent: read [history shape](./references/history-shape.md); choose state-specific disposition`"]
    Shape -->|No| StartReplay[Agent: start fresh replay in bound start worktree]; Shape -->|Failure or unobservable| Recover
    History --> HistoryResult{Supported state-specific disposition?}
    HistoryResult -->|Fresh start topology chosen| StartReplay; HistoryResult -->|Active commit: skip| Skip[Agent: skip only the classified active commit]
    HistoryResult -->|Active commit: preserve| Preserve[Agent: preserve classified active commit, then continue]; HistoryResult -->|No or conflicting| Decision
    HistoryResult -->|Failure or unobservable| Recover
    Stop -->|Unmerged entries| ReplayConflict; Stop -->|Topology, equivalent, or empty| History
    Stop -->|Metadata active; resolution staged or no conflict| ContinueReplay[Agent: continue the bound active replay]
    Stop -->|Other failure or unobservable| Recover
    StartReplay --> ReplayResult{Replay operation result?}; ContinueReplay --> ReplayResult; Skip --> ReplayResult; Preserve --> ReplayResult
    ReplayResult -->|Metadata/unmerged absent; R bound; result ref and bound goal relation match| Saved{Lifecycle saved entry?}
    ReplayResult -->|Unmerged entries| ReplayConflict["`Agent: read [conflict and ambiguity](./references/conflict-and-ambiguity.md); resolve replay intent`"]
    ReplayResult -->|Topology, equivalent, or empty| History; ReplayResult -->|Other failure or unobservable| Recover
    ReplayConflict --> ReplayIntent{Exactly one outcome preserves compatible intent, passes checks, and is staged?}
    ReplayIntent -->|Yes; active metadata unchanged| ContinueReplay; ReplayIntent -->|None or incompatible alternatives| Decision
    ReplayIntent -->|Failure or unobservable| Recover
    Abort --> AbortResult{Abort restoration predicate?}
    AbortResult -->|Metadata/unmerged absent; source or HEAD equals pre-replay OID| Saved; AbortResult -->|Failure or unobservable| Recover
    Saved -->|Verified none| FinishMode{Lifecycle outcome?}; Saved -->|Exact entry bound| Restore["`Agent: read [named stash](./references/named-stash.md); apply exact OID and preserve entry`"]
    Saved -->|Ambiguous or unobservable| Recover
    Restore --> RestoreResult{Apply result?}
    RestoreResult -->|Conflict-free| RestoreFinish[Agent: verify tree/checks; relist unique entry; drop selector; confirm absent]
    RestoreResult -->|Conflict; entry preserved| RestoreConflict["`Agent: read [conflict and ambiguity](./references/conflict-and-ambiguity.md); resolve restoration intent`"]
    RestoreResult -->|Other failure or unobservable| Recover
    RestoreConflict --> RestoreIntent{Exactly one outcome preserves compatible intent and passes checks?}
    RestoreIntent -->|Yes| RestoreFinish; RestoreIntent -->|None or incompatible alternatives| Decision
    RestoreIntent -->|Failure or unobservable| Recover
    RestoreFinish --> RestoreFinishResult{Checks zero, no unmerged entries, exact entry absent?}; RestoreFinishResult -->|Yes| FinishMode; RestoreFinishResult -->|No or unobservable| Recover
    FinishMode -->|Explicit abort| OrientAbort[Agent: inventory interactions; verify restored assumptions]
    FinishMode -->|Start or continue| Orient[Agent: bind committed result R; run repository and changed-interaction checks on its isolated tree]
    Orient --> VerifyResult{Checks pass on unchanged tracked tree of R; result ref equals R; no unmerged entries; intersections recorded?}
    VerifyResult -->|Yes| Publish{Publication bound and authorized?}
    VerifyResult -->|Failure attributable to replay; one correction within bound goal| Correct[Agent: apply and commit only the task-authorized replay correction]; VerifyResult -->|Not attributable, outside goal, missing intent, or alternatives| Decision
    VerifyResult -->|Failure or unobservable| Recover
    Correct --> CorrectResult{Correction committed within task authority?}
    CorrectResult -->|Yes; invalidate old evidence and rebind R| Orient; CorrectResult -->|Commit blocked or unauthorized| Decision; CorrectResult -->|Failure or unobservable| Recover
    Publish -->|No| Handoff; Publish -->|Yes| Remote["`Agent: read [publication](./references/publication.md); reconcile destination movement for this attempt`"]
    Remote --> RemoteResult{Remote stage result?}
    RemoteResult -->|Final fetch unchanged; exact lease current; R revalidated; result ref equals R| Push[Agent: push immutable R with the authorized exact lease]
    RemoteResult -->|Destination moved at final observation| Retry
    RemoteResult -->|Unexpected result-ref change| Decision
    RemoteResult -->|Conflict| RemoteConflict["`Agent: read [conflict and ambiguity](./references/conflict-and-ambiguity.md); resolve remote intent`"]
    RemoteResult -->|Failure or unobservable| Recover
    RemoteConflict --> RemoteIntent{Exactly one outcome preserves compatible intent and passes checks?}
    RemoteIntent -->|Exactly one| Retry; RemoteIntent -->|None or incompatible alternatives| Decision
    RemoteIntent -->|Failure or unobservable| Recover
    Push --> PushResult{Push result?}
    PushResult -->|Exit zero; post-fetch destination equals R| Handoff; PushResult -->|Lease rejected or destination moved| Retry{Attempts used below the cap in publication.md?}
    Retry -->|Yes| Remote; Retry -->|No| LeaseStop[Agent: stop external mutation; report each attempt, its rejection or movement, and the destination]
    PushResult -->|Other failure or unobservable| Recover
    NoChange --> Handoff; NoActive --> Handoff; Decision[Agent: stop mutation; report missing fact/alternatives and evidence] --> Handoff
    Recover["`Agent: read [active recovery](./references/active-rebase-recovery.md); stop mutation and collect report`"] --> Handoff
    OrientAbort --> Handoff; LeaseStop --> Handoff
    Handoff[Agent: carry the reached lifecycle path; check acquired worker obligation] --> Worker{Worker obligation?}
    Worker -->|None| Terminal{Reached lifecycle path and current evidence?}; Worker -->|Acquired| Deliver[Agent: deliver summary and resume/stop decision]
    Worker -->|Unknown or unobservable| Pending([Agent: stop mutation; preserve repo; report last worker state, handoff attempt, missing ack/observation; no completion])
    Deliver --> HandoffResult{Acknowledged summary/resume, or worker observed stopped with handoff?}
    HandoffResult -->|Yes| Terminal; HandoffResult -->|No or unobservable| Pending
    Terminal -->|Abort request and restored pre-state| Aborted([Aborted and restored])
    Terminal -->|Successful start/continue path; validation current for R; result ref and post-push destination equal R| Published([Published completion])
    Terminal -->|Successful start/continue path; validation current for R; result ref equals R; goal relation true; no publication| Local([Local completion])
    Terminal -->|No-change evidence complete| NoChangeDone([No change]); Terminal -->|Metadata absent; no mutation| NoActiveDone([No active rebase])
    Terminal -->|Decision report complete| Paused([Paused for decision]); Terminal -->|Stopped-state report complete| Stopped([Stopped with observed state])
    Terminal -->|Inconsistent or unobservable| Recover

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.