augments-labs
- 90 skills
- 0 followers
- 1 repo stars
- 6 days ago last updated
- ▌ Test Driven Development · augments-labs bundleDrives behavior changes through a failing test before implementation. Use when any behavior is added, changed, fixed, or preserved — features, fixes, refactors, migrations, generators, or configuration — even if the user never mentions tests or TDD. Skip throwaway spikes and nonbehavioral content or configuration.
- ▌ Containing An Incident · augments-labs bundleContains a failure that is reaching real users before diagnosing it. Use when there is an outage, a broken signup or checkout, a bad deploy, a spiking error rate, data corrupted or exposed, or a customer-visible regression, or when the user says it's down, customers are getting errors, something broke in production, or this started after the deploy. Skip a failing test, a bug caught in review, or a defect nobody has hit, and skip once impact has stopped.
- ▌ Dispatching Parallel Agents · augments-labs bundleSplits independent work across parallel agents, each with its own exclusive writes and resources. Use when tasks can run concurrently with no dependence on another worker's output, such as separate files, services, or investigations; shared reads of frozen inputs are fine. Skip coupled writes, shared runtime state, dependent outputs, or work quicker to do inline.
- ▌ Refactor Architecture · augments-labs bundleRedesigns the structure of existing code when that structure makes change expensive: tangled boundaries, logic in the wrong layer, one small change touching many files. Use when the structure of existing code is what makes change expensive and needs redesigning, or when the user says this codebase is a mess, everything imports everything, or every feature makes the next one slower. Skip designing a new system, quick local cleanups, and read-only audits.
- ▌ Visual UI Verification · augments-labs bundleVerifies a running interface visually across states, viewports, themes, and input paths. Use when a UI-bearing change is about to be called visually correct, done, or ready to ship, when acceptance depends on how a running GUI or TUI looks or responds, or on does this look right, is the UI done, or check the screen. Skip isolated widget snapshots, nonvisual behavior, and open design decisions.
- ▌ Requesting Code Review · augments-labs bundleGets an independent review of a frozen code state before completion, push, or merge. Use when a change reaches a done or integration boundary, when review of a frozen state is requested, or when the user says review this, is this ready to merge, or take a look before I push. Skip unfinished checkpoints unless requested, keep/discard/close/reopen transitions, and plan documents, which `reviewing-plans` reviews.
- ▌ Executing Plans · augments-labs bundleRuns an approved plan task by task in this session, through each task's evaluator. Use when the user asks to execute, continue, or resume work governed by an approved plan directory in inline mode, including one remaining task or a reply choosing inline mode. Skip a standalone task with no plan directory, and a plan approved in delegated mode.
- ▌ Verification Strategy · augments-labs bundleDesigns or repairs a project's correctness battery: which checks catch which defects and when they run. Use when a project's assurance is absent, stale, or unfalsifiable before high-risk work, when risk changes, after an escaped defect or a hollow gate, or when the user asks how to test this project, says the tests don't catch anything, or asks what CI should run. Skip a bounded feature, and skip writing or running an already-defined gate.
- ▌ Receiving Code Review · augments-labsVerifies review feedback on its merits before responding to it or editing anything. Use when identifiable human or automated review findings arrive, including stale, ambiguous, or conflicting feedback. Skip a claimed review with no report and raw build or test failures that need technical diagnosis.
- ▌ Architecture Decisions · augments-labs bundleRecords a significant, hard-to-reverse technical decision and the options weighed. Use when a choice such as a datastore, sync versus async, a framework, a public contract, or an auth or security model is being weighed or has just been settled, when the user asks whether to use X or Y, when a choice is made in passing during discussion, or when an old decision is revisited. Skip easily reversible choices.
- ▌ Finishing A Branch · augments-labs bundleTakes a completed branch through the transition the user chooses. Use when implementation is complete and gates are green with an unsettled branch choice, or when the user asks to push, integrate, keep, discard, close, or reopen the branch, what to do with this branch, to merge it, to open a PR, or whether we are done here. Skip mid-development checkpoints and already-settled branch states.
- ▌ Complexity Audit · augments-labs bundleAudits existing code for accidental complexity: abstraction nothing needs, ownership it should not hold, flexibility nobody uses, or custom machinery a library already provides. Use when existing code should be examined for accidental complexity, or when the user asks whether this is over-engineered, why it is so complicated, or whether all of it is still needed. Skip implementation choices, review of a frozen change, and structural work already approved.
- ▌ Release Readiness · augments-labs bundleDecides whether an integrated candidate is safe to promote to a release, deployment, or publication. Use when the user says ship it, asks to cut a release or publish the package, or asks whether this is safe to deploy, and before a first release, a version bump, a package publish, a production deploy, or a rollout stage expansion. Skip internal work with no release surface.
- ▌ Feasibility Check · augments-labs bundleChecks whether an initiative can be delivered before committing to it: technical, delivery, budget, operational, security, data, and dependency risk. Use when an initiative's feasibility is uncertain, including proven technology under uncertain delivery constraints. Skip only when an existing feasibility section already fills every dimension row from evidence and no killer risk's confidence is unknown.
- ▌ System Architecture · augments-labs bundleDesigns the target system for approved requirements: components, boundaries, data flow, failure and recovery, and which seams are worth their cost. Use when approved requirements need a system design before planning or implementation, or when the user asks how to structure this or what the moving pieces are. Skip the transition from an existing system and a small feature whose structure fits inside its own task or plan.
- ▌ Using Git Worktrees · augments-labs bundleCreates an isolated git worktree for a task and checkpoints work there. Use when a repository is about to be edited for the first time in a task, when work needs an isolated checkout and its own resources, or at a local task checkpoint. Reuse a linked worktree only when this task or the harness owns it; there, skip creating a workspace but keep checkpoint discipline. Skip read-only work.
- ▌ Migration Strategy · augments-labs bundlePlans how a rewrite or migration preserves behavior through cutover and recovery. Use when a rewrite, migration, or broad preservation-sensitive transformation is about to be planned or implemented and its behavior, cutover, or recovery cannot be established by ordinary review and gates. Skip bounded changes whose diff, behavior, and recovery remain directly reviewable.
- ▌ Viewing Artifacts · augments-labs bundleShows the state of the artifact trail — briefs, specs, designs, plans, execution — at a glance instead of file by file. Use when the user asks where the project stands, what needs attention, whether the plan is still in sync with the spec, to see the state of specs and plans, or to visualize the trail. Skip when one artifact must be read, written, or edited.
- ▌ Using Sdlc Skills · augments-labs bundleRoutes each task to the SDLC skills it needs before work begins. Use when a task opens, resumes, or is handed off, and at every material state change — a phase ends, a decision returns, feedback arrives — before any answer, question, exploration, or tool call, as with build X, fix this, is it done, or plan this.
- ▌ Security Audits · augments-labs bundleAudits what an attacker can make a change do, from attacker-controlled input to its effect. Use when a change touches a trust boundary, attacker-controlled input, authentication or authorization, secrets, sensitive data, isolation, a dependency, build or deploy exposure, availability, sessions, or cryptography, including anything touching login, tokens, permissions, uploads, or user input reaching a query or a shell. Skip only when no security surface changed.
- ▌ Post Mortem · augments-labs bundleExplains why safeguards missed a production escape, incident, loss, or failed work cycle, or why its impact grew, and what prevents it happening again. Use when a production escape, incident, material loss, or failed work cycle has been contained, or when the user asks how this reached production or what prevents another failed handoff. Establish any technical cause and immediate containment first; a process-only failure needs its event evidence. Skip ordinary bugs and ongoing uncontained impact.
- ▌ Coding Standards · augments-labs bundleSettles the conventions agents and humans both follow in a project: domain vocabulary, naming, patterns to reach for, and things to never do. Use when a project needs its conventions settled, once per project or when they have drifted, or when the user asks for a style guide or house rules, says the codebase is inconsistent, or asks how something should be named. Skip when clear standards already exist and are followed, and skip proving the code correct.
- ▌ Define Goals · augments-labs bundlePins down what a new project or initiative is for: the objective, the stakeholders, how success is measured, and the guardrails. Use when a new project or initiative starts before scoping or building, or when the user wants to build X with no stated objective, or when the objective is vague or missing. Skip a single feature, and skip while a decision or approval reply is pending.
- ▌ Debugging · augments-labs bundleFinds the cause of a bug before any fix is proposed or applied. Use when a bug, test failure, flaky or intermittent result, or unexpected behavior has an unknown technical cause, or when the user says it's broken, this doesn't work, or why is it doing that. Skip a one-line error whose cause and complete effect are directly visible, and skip explaining how a known, contained production failure escaped its safeguards.
- ▌ Writing Skills · augments-labs bundleWrites and edits the skills in this library: the SKILL.md format, descriptions, Gotchas, support files, and the proof that a skill changes what an agent does. Use when adding a skill, changing or fixing one, or when a skill is not triggering. Skip using a skill.
- ▌ Writing Plans · augments-labs bundleWrites an executable plan: ordered tasks, their dependencies, and how each one is known to be done. Use when approved requirements or a clearly multi-step task need a per-task plan before implementation, or when the user says break this down, lay out the steps first, or this is too big to do in one go. Skip single-step or trivial work.
- ▌ UI UX Design · augments-labs bundleDesigns a user interface's flow, states, hierarchy, responsive behaviour, accessibility, content, and visual direction before it is built. Use when a new or revised interface still has open design decisions, or when the user says design this screen, asks what it should look like, or how the user gets through it. Skip backend-only work and cosmetic edits whose direction is already decided.
- ▌ Prototyping · augments-labsBuilds a throwaway prototype to answer an uncertain design or feasibility question faster than arguing about it. Use when the user says let's just try it and see, spike this, or quick and dirty, I'll throw it away, or when a tricky bit of logic, a library's real behaviour, or whether a UI can be built, render, or perform as designed is genuinely uncertain and cheaper to settle with a throwaway. Skip when the answer is already known, and send a layout or variant choice to ui-ux-design.
- ▌ Data Model · augments-labs bundleDefines domain concepts, relationships, state transitions, and invariants before implementation. Use when any of them needs to be introduced, changed, or removed, including stateless rules and pricing engines. Skip when those semantics stay unchanged or are already settled for this task.
- ▌ Handoff · augments-labs bundleWrites a handoff so a fresh session or another agent can continue unfinished work. Use when a session is ending mid-work, the context is getting full, the user is heading out, or someone else is taking over. Skip resuming from a handoff an earlier session wrote: reading it needs no skill. Skip finished tasks and bounded worker assignments whose coordinator stays responsible.
- ▌ Yagni · augments-labs bundleChallenges scope so only what the task needs gets built, and built completely. Use when code, files, flags, or dependencies are being removed as unused, dead, or legacy; when scope drifts toward speculative or incomplete delivery; when a proposal needs a strict pre-edit challenge; or when anyone says we might need this later or let's make it configurable. Skip throwaway spikes and content or configuration with no behavior.
- ▌ Adding Observability · augments-labsInstruments a change or service with logs, metrics, traces, and symptom alerts. Use when adding logs, metrics, traces, or an alert, or when the user says add observability, add monitoring, instrument this, we have no visibility, or we cannot tell whether it is working. Skip rollout threshold signals, which rollout-controls owns, and reading existing signals to find a cause, which debugging owns.
- ▌ Eliminating Corrections · augments-labsTurns a correction a human made to the agent's work into the strongest proportionate way of making that class of mistake impossible, automatic, or caught. Use when the user corrects what the agent did or how it did it, says no, don't do that, that's wrong, it should be like this, I told you this before, or stop doing this, reverts an edit the agent made, repeats a review comment on a later change, or asks how to stop this kind of mistake from happening again. Skip a bug whose cause is unknown, a failure that reached production, and a one-off preference that cannot recur.
- ▌ Changing A Live Schema · augments-labs bundlePhases a live schema change so the store serves both code versions through deploy and rollback. Use when adding, renaming, dropping, or retyping a column or field, adding an index or constraint, splitting a table, changing a stored document shape, or backfilling data on a live store. Skip stores nothing live reads yet or with recorded accepted downtime, and whole-store migrations, which migration-strategy owns.
- ▌ Verification Before Completion · augments-labs bundleRuns the checks and reads their output before any claim that work is done. Use when work is about to be called complete, fixed, passing, done, or satisfactory, before any commit or PR, when evidence may be stale, partial, or bound to another state, or before saying that should do it, it's working now, or all set, even if no formal claim is made.
- ▌ Subagent Driven Development · augments-labs bundleRuns an approved plan's tasks through fresh subagents with controller-owned review and gates. Use when an approved plan is built by dispatched workers rather than in this session, when the user asks for subagents or workers to build it, or when the user chooses delegated mode. Skip a lone task with no plan directory.
- ▌ Mapping The Codebase · augments-labsMaps how a region of code fits together — its callers, owners, and boundaries — before it is changed. Use when code whose structure and callers have not been established for this task is about to be planned, changed, or debugged, or when asked how a change travels through the codebase. Skip only when a current boundary record covers the region and its inputs.
- ▌ Working From A Ticket · augments-labs bundleCaptures a ticket from an issue tracker as the definition of the work: its text and acceptance criteria as read, and what it leaves unclear, so the plan and the tests bind to a recorded text. Use when the user hands over a ticket, issue, story, or work item as the task, whether pasted, linked, or named by its identifier, and again before ticket-defined work is called done. Skip work the user describes in their own words with no ticket behind it.
- ▌ Deprecating An API · augments-labs bundleRetires a public function, method, class, endpoint, response field, CLI flag, config key, event, or message type without breaking callers. Use when removing, replacing, renaming, or sunsetting something others call, setting a removal date or sunset, or when the user says deprecate this, retire this, or stop supporting this. Skip interfaces whose callers are all inside this repository and stored data shapes, which changing-a-live-schema owns.
- ▌ Rollout Controls · augments-labs bundleAdds the feature flag or kill switch and threshold signals a controlled rollout needs. Use when release policy, the plan, or the user calls for a feature flag, kill switch, canary, or staged rollout, or when release readiness blocks on missing rollout policy or rollout signals. Skip changes nobody asked to stage and migrations or high-risk transformations whose own contract owns rollout.
- ▌ Large Scale Changes · augments-labs bundleApplies a mechanical change across many files, packages, or owners with a script and per-owner shards. Use when writing or running a codemod, or when the user says migrate every call site, rename this everywhere, bump X across all packages, roll out this lint rule or config everywhere, or split this into PRs per owner. Skip changes that fit one reviewable PR and structural redesigns, which refactor-architecture owns.
- ▌ Diagnosing A Session · augments-labs bundleReconstructs what an agent session did from its recorded transcript and shows where it went wrong. Use when the user asks why a session did something, what happened earlier in a run, why an agent skipped a step, took an action nobody asked for, or lost track of the task, or wants a session transcript or agent run inspected. Skip defects in the software being built and failures still reaching users.
- ▌ Writing Specs · augments-labs bundleWrites a specification: detailed requirements and acceptance criteria for settled intent, or a revision to an existing spec. Use when settled intent needs detailed requirements and acceptance criteria before building, including a revision to an existing spec, or when the user asks for a spec, requirements, or acceptance criteria. Skip when current requirements already cover the work, when choosing implementation details, or while a material decision about what the user wants is still pending.
- ▌ Reviewing Plans · augments-labs bundleGets an independent review of one exact plan version before the plan is presented for approval. Use when a written plan's route is high-risk and it is about to be shown for approval, or when the user says review this plan, is the plan sound, or check the plan before I approve it. Skip a plan whose files are not written yet, and skip editing the plan.
- ▌ Brainstorming · augments-labs bundleGenerates several distinct approaches to a settled goal, each under a different design constraint and checked against the same facts about the current code, then compares them so one is chosen before anything is specified or built. Use when the goal is clear but the approach is open, when the user says brainstorm, what are our options, how should we do this, which approach, or compare alternatives, or before a spec or design when only one approach has been considered. Skip when the approach is already fixed, when the goal itself is open, and while a decision reply is pending.
- ▌ Reviewing A Pr · augments-labs bundleReviews a pull request someone else wrote. Use when a teammate's or contributor's PR is assigned to you or waiting on your review, or when the user says review PR 42, look at this pull request, what do you think of this contribution, or should we accept this. Skip changes you wrote yourself, which `requesting-code-review` reviews, and feedback arriving on your own change, which `receiving-code-review` verifies.
- ▌ Runtime Evidence · augments-labs bundleAnswers a question about how existing code behaves by running it, and backs a claim about speed, memory, or size with a controlled measurement. Use when asked what a function returns, whether a case is handled, which branch or code path runs, why an output appears, what a setting resolves to at runtime, or how an installed dependency behaves, and before saying code is faster, lighter, or smaller. Skip when the question is about intent, design, or history, when something is broken and the cause is unknown, or when the code does not exist yet.
- ▌ Clarifying Intent · augments-labs bundleSettles ambiguous user intent or an unanswered decision before work proceeds. Use when a request has several plausible readings that change the work, or when a pending decision receives praise, constraints, silence, or a partial reply instead of an explicit choice or cancellation, even if nobody asks for clarification. Skip inputs another skill's own procedure asks for.
- ▌ Scoping · augments-labs bundleDefines a project's delivery boundary: what is in, what is out, and what is sufficient for the goal. Use when a project's goals are set and its scope needs defining or revising, including scope growth during delivery. Skip single-feature requirements, unresolved intent, and speculative additions at the implementation level.
- ▌ Yagni 2 · augments-labs bundleUse when code, files, flags, or dependencies are being removed as unused, dead, or legacy; when behavior-affecting implementation or configuration is being written or proposed; when scope drifts toward speculative or incomplete delivery; or when a proposal needs a strict pre-edit challenge. Fires on we might need this later and let's make it configurable, even if nobody says scope or YAGNI. Skip throwaway spikes and nonbehavioral content or configuration.
- ▌ Handoff 2 · augments-labs bundleUse when a session is ending or work is passing to a fresh session or another agent. Fires on I'm heading out, wrapping up for the day, context is getting full, and someone else is taking this over, even if nobody says handoff. Skip a finished, self-contained task that needs no continuation.
- ▌ Spec It 2 · augments-labs bundleUse when settled intent for a goal or feature needs detailed requirements before design — what it must do, how each requirement is verified, and the assumptions and risks behind it. Fires on write up what this needs to do and what are the acceptance criteria, even if nobody says spec or requirements. Skip how it is built, skip when requirements already exist, and skip while a decision or approval reply is pending.
- ▌ Scope It 2 · augments-labs bundleUse after the goals are set and before design, to draw a project's boundary: what is in, what is explicitly out, and the smallest cut that still meets the goal. Fires on what goes in v1, what should we cut, and this is getting too big, even if nobody says scope. Skip a single feature, unresolved ambiguity about intent, and detailed feature requirements.
- ▌ Zoom Out 2 · augments-labsUse before working in a region of the codebase whose structure you have not established — its modules, their callers, and the project's own vocabulary. Fires whenever a request names an unfamiliar file, module, or area, and on any structural question — what reaches this, what it reaches, how a change travels through it — even if nobody asks to be oriented. Skip when the area is already understood.
- ▌ Debugging 2 · augments-labs bundleUse before proposing or applying a fix to any bug, test failure, flaky or intermittent result, or unexpected behavior whose technical cause is unknown. Fires on it's broken, this doesn't work, and why is it doing that, even if nobody says debug. Skip a one-line error whose cause and complete effect are directly visible, and skip explaining how a known, contained production failure escaped its safeguards.
- ▌ Data Model 2 · augments-labs bundleUse when the domain needs modeling before design or code — the concepts a system stores or computes over: entities, relationships, state transitions, and the invariants that keep them true. Fires for a stateless engine (pricing, rules, workflow) as much as for schema design, and on what are the core objects here, even if nobody says data model or schema. Skip a feature that adds no domain concepts.
- ▌ Post Mortem 2 · augments-labs bundleUse after a production escape, late defect, data loss, outage, security incident, or badly failed work cycle, once the technical cause and containment are known and the open question is why the safeguards missed it or why the impact grew. Fires on how did this reach production, why didn't we catch this, and what do we change so it doesn't happen again, even if nobody says post-mortem. Skip while the technical cause is still unknown, and skip ordinary bugs.
- ▌ Prototyping 2 · augments-labsUse when a design or feasibility question is genuinely uncertain and cheaper to answer by building a throwaway than by arguing about it — a tricky bit of logic, a layout choice, a library's real behaviour. Fires on let's just try it and see, spike this, and quick and dirty, I'll throw it away, even if nobody says prototype. Skip when the answer is already known.
- ▌ Define Goals 2 · augments-labs bundleUse at the start of a new project or initiative, before scoping or building, to pin down what it is for: the objective, the stakeholders, how success is measured, and the guardrails. Fires on we want to build X with no stated objective, and on a vague or missing objective, even if nobody says goals. Skip a single feature, and skip while a decision or approval reply is pending.
- ▌ Interview Me 2 · augments-labs bundleUse when material intent has several plausible readings that no owning phase skill can settle, or when a pending material decision got neither an explicit answer nor an explicit cancellation — praise, constraints, silence, or a partial reply instead. Fires on a reply that approves of the work without choosing among the options put to the user, even if nobody asks for questions or clarification. Skip inputs an owning skill's own procedure elicits.
- ▌ UI UX Design 2 · augments-labs bundleUse when a new or revised user interface still has open decisions about flow, state, hierarchy, responsive behaviour, accessibility, content, or visual direction, before it is implemented. Fires on design this screen, what should this look like, and how does the user get through this, even if nobody says UX or design. Skip backend-only work and cosmetic edits whose direction is already decided.
- ▌ Writing Plans 2 · augments-labs bundleUse when approved requirements or a clearly multi-step task need an executable, per-task plan before implementation — ordered tasks, their dependencies, and how each one is known to be done. Fires on break this down, lay out the steps first, and this is too big to do in one go, even if nobody says plan. Skip single-step or trivial work.
- ▌ Writing Skills 2 · augments-labs bundleUse when creating or editing a skill in this library — its format, its sibling files, and how to prove it works. Fires on add a skill for X and this skill isn't triggering, even if nobody says authoring. Skip using a skill.
- ▌ Executing Plans 2 · augments-labs bundleUse when asked to execute, run, continue, or resume a multi-task plan directory, and immediately after the user selects an execution mode for an approved plan — even through a terse reply such as inline, delegated, option 1, or option 2. Fires on pick up where we left off when a plan directory holds the work. Skip a single task.
- ▌ Security Audits 2 · augments-labs bundleUse when a change touches a trust boundary, attacker-controlled input, authentication or authorization, secrets, sensitive data, isolation, a dependency, build or deploy exposure, availability, sessions, or cryptography. Fires on anything touching login, tokens, permissions, uploads, or user input reaching a query or a shell, even if nobody says security. Skip only when no security surface changed.
- ▌ Coding Standards 2 · augments-labs bundleUse once per project, or when conventions have drifted, to settle the conventions agents and humans both follow — domain vocabulary, naming, patterns to reach for, things to never do. Fires on a request for a style guide or house rules, on the codebase is inconsistent, and on how should we name this, even if nobody says standards or conventions. Skip when clear standards already exist and are followed, and skip proving the code correct.
- ▌ Complexity Audit 2 · augments-labs bundleUse when existing code should be examined for accidental complexity — abstraction nothing needs, ownership it should not hold, flexibility nobody uses, or custom machinery a library already provides. Fires on is this over-engineered, why is this so complicated, and do we still need all of this, even if nobody says complexity or audit. Skip implementation choices, review of an exact candidate, and structural work already approved.
- ▌ Feasibility Check 2 · augments-labs bundleUse before an accountable owner commits to a project or initiative, when whether it can be done under real technical, delivery, operational, security, data, or dependency constraints is still uncertain. Fires on can we actually build this, is this realistic by the deadline, and what would it take, even if nobody says feasibility. Skip proven or trivially reversible approaches.
- ▌ Release Readiness 2 · augments-labs bundleUse when an integrated candidate is about to be promoted to a release, deployment, or publication — the first release of a project, a version bump, a package publish, a production deploy, or a rollout stage expansion. Fires on ship it, cut a release, publish the package, and is this safe to deploy, even if nobody says release. Skip internal work with no release surface.
- ▌ Using Sdlc Skills 2 · augments-labs bundleUse at every task opening, resume, or handoff, and again at every material change of state — a phase ends, a decision returns, feedback arrives — before any answer, question, exploration, or tool call that begins the work. Fires on any opening — build X, fix this, it's down, is it done, review this, plan this — even if nobody mentions skills or process.
- ▌ Viewing Artifacts 2 · augments-labs bundleUse when the state of the artifact trail — briefs, specs, designs, plans, execution — needs to be seen at a glance instead of read file by file. Fires on show me the state of my specs and plans, where does my project stand, what needs attention, is my plan still in sync with the spec, and visualize the trail, even if nobody says artifact or viewer. Skip when one artifact must be read, written, or edited.
- ▌ Finishing A Branch 2 · augments-labs bundleUse when implementation is complete and its gates are green, and the work now needs an integration decision — push, open, update or merge a PR, integrate locally, keep, or discard — and whenever the user explicitly chooses to keep, discard, close, or reopen a branch, workspace, or PR. Fires on ship it, what do we do with this branch, and are we done here, even if nobody names a git operation. Skip mid-development checkpoints.
- ▌ Migration Strategy 2 · augments-labs bundleUse before planning or implementing a high-risk rewrite, migration, generated conversion, or wide behavior-preserving transformation — one whose output cannot be reviewed line by line, crosses ownership boundaries, or can fail at the platform, build, runtime, data, or cutover layer. Fires on port this to X, move us off Y, and regenerate this from Z, even if nobody says migration. Skip bounded changes.
- ▌ System Architecture 2 · augments-labs bundleUse when approved requirements need a target system design before planning or implementation: components, boundaries, data flow, failure and recovery, and which seams are worth their cost. Fires on how should we structure this and what are the moving pieces, even if nobody says architecture. Skip the transition from an existing system, and skip a small feature whose structure fits inside its own task or plan.
- ▌ Using Git Worktrees 2 · augments-labs bundleUse before the first repository edit — a feature, fix, refactor, plan step, or dispatched agent work — or whenever work needs isolation from the current checkout. Fires on start on this ticket and let's build X inside a repository, even if nobody says branch or worktree. Skip read-only work, and skip when the current checkout is already dedicated to this task.
- ▌ Verifying Completion 2 · augments-labs bundleUse before any claim that work is complete, fixed, passing, done, or satisfactory, before any commit or PR, and when evidence may be stale, partial, or bound to another state. Fires on that should do it, it's working now, and all set, even if no formal claim is made.
- ▌ Receiving Code Review 2 · augments-labsUse when identifiable review feedback arrives, from a human or another agent, before responding, editing, or resolving any finding, and when feedback is stale, ambiguous, or conflicting. Fires on pasted review comments, a reviewer's list of concerns, and here is what CI flagged, even if nobody says code review. Skip a claimed or in-flight review with no returned report.
- ▌ Refactor Architecture 2 · augments-labs bundleUse when the structure of existing code is what makes change expensive and that structure itself needs redesigning — tangled boundaries, logic in the wrong layer, one small change touching many files. Fires on this codebase is a mess, everything imports everything, and every feature makes the next one slower, even if nobody says refactor or architecture. Skip designing a new system, quick local cleanups, and read-only audits.
- ▌ Verification Strategy 2 · augments-labs bundleUse to establish or repair a project's correctness battery, before high-risk work whose assurance is absent, stale, or unfalsifiable, and again on changed risk, an escaped defect, or a hollow gate. Fires on how should we test this project, our tests don't catch anything, and what should CI run, even if nobody says strategy. Skip a bounded feature, and skip writing or running an already-defined gate.
- ▌ Architecture Decisions 2 · augments-labs bundleUse when a significant, hard-to-reverse technical choice is being weighed or has just been settled — a datastore, sync vs async, a framework, a public contract, an auth or security model. Fires on should we use X or Y, on a choice made in passing during discussion, and on a request to revisit an old one, even if nobody says ADR or decision record. Skip easily-reversible choices.
- ▌ Containing An Incident 2 · augments-labs bundleUse the moment a failure is reaching real users — an outage, a broken signup or checkout, a bad deploy, a spiking error rate, data corrupted or exposed, a customer-visible regression. Fires on it's down, customers are getting errors, something broke in production, and this started after the deploy, even if nobody says incident or outage. Skip a failing test, a bug caught in review, or a defect nobody has hit, and skip once impact has stopped.
- ▌ Requesting Code Review 2 · augments-labs bundleUse when an exact candidate reaches a done or integration boundary, or an independent review of one frozen state is explicitly requested. Fires on review this, is this ready to merge, and take a look before I push, even if nobody says code review. Skip an unfinished reversible checkpoint unless review was explicitly requested, and skip an explicit keep, discard, close, or reopen.
- ▌ Visual UI Verification 2 · augments-labs bundleUse before declaring any page, screen, view, or other UI-bearing candidate visually correct, done, or ready to ship, and when acceptance or release depends on how a running GUI or TUI looks or responds across states, viewports, themes, or input paths. Fires on does this look right, is the UI done, and check the screen, even if nobody asks for a visual check. Skip isolated widget or snapshot assertions, nonvisual behavior, and open design decisions.
- ▌ Test Driven Development 2 · augments-labs bundleUse before behavior-affecting implementation — features, fixes, refactors, migrations, generators, or configuration, including behavior meant to be preserved. Fires on any request to add, change, or fix behavior, even if the user names only the feature or the bug and never mentions tests or TDD. Skip throwaway spikes and nonbehavioral content or configuration.
- ▌ Dispatching Parallel Agents 2 · augments-labs bundleUse when two or more work items can run concurrently with disjoint file ownership, mutable state, and outputs — unrelated bugs, independent features, parallel research. Fires on do these at the same time and can several agents work on this, even if nobody says parallel or concurrent. Skip shared generators, manifests, or runtime state, dependent outputs, and work quicker to do inline.
- ▌ Zoom Out · augments-labsUse before working in a region of the codebase whose structure you have not established — its modules, their callers, and the project's own vocabulary. Fires whenever a request names an unfamiliar file, module, or area, and on any structural question — what reaches this, what it reaches, how a change travels through it — even if nobody asks to be oriented. Skip when the area is already understood.
- ▌ Spec It · augments-labs bundleUse when settled intent for a goal or feature needs detailed requirements before design — what it must do, how each requirement is verified, and the assumptions and risks behind it. Fires on write up what this needs to do and what are the acceptance criteria, even if nobody says spec or requirements. Skip how it is built, skip when requirements already exist, and skip while a decision or approval reply is pending.
- ▌ Scope It · augments-labs bundleUse after the goals are set and before design, to draw a project's boundary: what is in, what is explicitly out, and the smallest cut that still meets the goal. Fires on what goes in v1, what should we cut, and this is getting too big, even if nobody says scope. Skip a single feature, unresolved ambiguity about intent, and detailed feature requirements.
- ▌ Interview Me · augments-labs bundleUse when material intent has several plausible readings that no owning phase skill can settle, or when a pending material decision got neither an explicit answer nor an explicit cancellation — praise, constraints, silence, or a partial reply instead. Fires on a reply that approves of the work without choosing among the options put to the user, even if nobody asks for questions or clarification. Skip inputs an owning skill's own procedure elicits.
- ▌ Verifying Completion · augments-labs bundleUse before any claim that work is complete, fixed, passing, done, or satisfactory, before any commit or PR, and when evidence may be stale, partial, or bound to another state. Fires on that should do it, it's working now, and all set, even if no formal claim is made.