Skill Registry

Use when the user wants lint, format, and type errors fixed in the current changes. Reads CLAUDE.md for the repo's lint/format/type-check commands, runs each, and fixes only style/format/type issues — no behavior changes.

Use when the user wants a PR description generated for the current branch. Reads commit history, file changes, and CLAUDE.md, then writes a Summary / Changes / Test Plan / Notes block to pr-description.md.

Use when the user wants to restructure code while preserving behavior exactly. Establishes a passing test baseline first, then makes incremental moves that each leave the suite green. Refuses to change behavior and structure in the same step.

Use when the user wants to cut a release — bump the version, update the changelog from conventional commits, and tag. Detects where the version lives, derives the next version from the commits since the last tag, and stages the release locally; it does not push or publish unless explicitly asked.

Use right before declaring an implementation done — a last-pass review of your OWN uncommitted change against a fixed checklist (reuse, stdlib, comments, dead code, tests, scope). Catches the things that make a diff read as AI-written before a human ever sees it. Reviews the current diff; it does not write new features.

Use when the user wants tests written for current changes (uncommitted diff or recent feature). Matches the repo's existing test framework, fixtures, and assertion style. Covers happy path, edge cases, and error paths without over-mocking.

Vast.ai CLI to manage GPU instances, volumes, serverless endpoints, and billing.

Create a new Langfuse integration page in the langfuse-docs repo. Use this skill whenever the user wants to add, create, draft, or scaffold an integration page, cookbook, or docs page for a new tool/framework/model-provider/gateway in Langfuse — triggers include "new integration", "integration page", "docs page for <X>", "cookbook for <X>", "add <X> to langfuse docs", or any request that results in a new `cookbook/integration_*.ipynb`. Also use when the user pastes working integration code, a link to a partner's docs, or rough notes and wants them turned into the standard Langfuse integration notebook. The skill produces a correctly formatted Jupyter notebook, updates `cookbook/_routes.json`, and tries to fetch the partner logo into `public/images/integrations/`.

An advanced, autonomous AI agent skill designed to execute complex, multi-step, and long-horizon tasks with high reliability and minimal human interruption.

Use when a task needs the judgment of a Ranch and Aquacultural Farmworker — scoring livestock body condition and lameness, responding to a dissolved-oxygen crash in an aquaculture pond, running low-stress animal handling for vet work or transport, applying herd or pond biosecurity and quarantine protocol, or diagnosing a feed-conversion-ratio drift before it becomes a mortality event.

Use when a task needs the judgment of a Petroleum Pump System Operator, Refinery Operator, or Gauger — diagnosing a product quality failure when every individual temperature and pressure reading looks within range, investigating a relief valve lift's upstream cause, applying temperature/API gravity correction to a tank gauge reading for custody transfer, or deciding how to treat a deviation during startup versus steady-state operation.

Use when a task needs the judgment of a Registered Nurse — recognizing early clinical deterioration from vital-sign trends, prioritizing a multi-patient assignment by acuity, escalating a change in patient condition to a physician (SBAR), applying medication-safety checks under time pressure, or delegating tasks to techs/CNAs within their scope.

Use when a task needs the judgment of a Retail Loss Prevention Specialist — deciding whether a floor observation clears the bar for a shoplifting stop, running plainclothes surveillance across a heat-mapped shrink zone, executing a stop safely without exceeding no-pursuit/no-contact policy, triaging a self-checkout scan-avoidance case, or writing the incident report and chain-of-custody log that a case will stand or fall on.

Use when a task needs the judgment of a security and fire alarm systems installer — laying out detector/notification-appliance placement against NFPA 72, choosing Class A vs Class B circuit topology, sizing candela on notification appliances, prepping for an AHJ acceptance test, or diagnosing a false-alarm pattern on an existing system.

Use when a task needs the judgment of a (physical/corporate) Security Manager — designing physical access controls, assessing a facility or event's security risk, responding to a security incident, or deciding how to allocate a security budget across prevention, detection, and response. Distinct from a cybersecurity/information-security role — this one covers physical and personnel security.

Use when a task needs the judgment of a subway or streetcar operator — deciding whether to hold at a stop or run to close a headway gap without violating minimum train separation, evaluating a door-obstruction sensor event before recycling or manually overriding the door, judging an ATC/ATO-to-manual mode transition and its speed restriction, planning a tunnel emergency-evacuation communication sequence, or verifying platform gap-filler deployment at a curved station before releasing doors.

Use when a task needs food-preparation-worker judgment — building tomorrow's prep list from a covers forecast, converting a yield percentage into an order quantity, labeling and rotating stock under FIFO, sequencing cuts to minimize danger-zone dwell time, or deciding whether a batch is still safe to use.

Use when a task needs the judgment of a Food Scientist — designing an accelerated shelf-life study, running a HACCP hazard analysis on a process step, planning a sensory-evaluation protocol (triangle test, descriptive panel), or reformulating a product for cost/nutrition targets without breaking stability or safety.

Use when a task needs the judgment of a Geoscientist — correlating well logs across a stratigraphic section, estimating hydrocarbon or mineral reserves volumetrically, assessing geologic-hazard risk for a site, or interpreting subsurface structure from seismic or core data.

Use when a task needs the judgment of a Light Truck Driver — checking whether a vehicle reassignment crosses the 26,001 lb CDL threshold, deciding whether to skip a backing safety check to hit a delivery-window cutoff, classifying a proof-of-delivery or signature exception, planning stop-sequencing against a shift's on-duty time budget, or responding to a residential hazard (loose dog, blocked driveway, unsafe backing approach) mid-route.

Use when a task needs the judgment of a Machine Feeder and Offbearer — timing a manual feed/discharge task against a machine's actual cycle time to find a hidden bottleneck, verifying a two-hand control or presence-sensing device's minimum safety distance after a tooling or stock changeover, deciding whether a point-of-operation exposure needs a pullback device versus a restraint device versus presence-sensing, judging whether a light-curtain nuisance trip is an ergonomic fix or a defeat-in-progress, or holding a station when pace pressure has pushed hands closer to the point of operation than the safeguard's timing allows.

Use when a task needs the judgment of an Information Security Analyst — triaging and investigating a security alert or suspected intrusion, designing or tuning detection rules and access controls, running incident response through containment/eradication/recovery, prioritizing a vulnerability-patching backlog, or reconstructing an attack timeline across compromised hosts.

Use when a task needs nanosystems-engineer judgment — choosing a nanofabrication process for a target production volume, diagnosing an ALD or lithography conformality/overlay failure from metrology data, designing nanomaterial exposure controls with no regulatory PEL to lean on, or judging whether a nanomaterial or nanoscale device claim is actually production-ready versus still lab-stage.

Use when a task needs the judgment of a Nuclear Power Reactor Operator — planning a power change around the xenon-135 transient's delayed peak rather than just the immediate reactivity effect, verifying shutdown margin against the worst-case stuck-rod condition rather than observed rod positions, accounting for a control rod's position-dependent worth, or deciding whether to deviate from an approved procedure's step sequence.