AI & ML

kanevry Skill

Use this skill to run an honest session-process evaluation (Standard v1, aiat-llm-eval/1.0) — score the last completed orchestrator session against the pre-registered rubric-v1 dimensions, run /eval, evaluate this session, produce an eval report, or re-verify a stored eval run for reproducibility. Deterministic-first with an optional advisory LLM judge; never produces a global score.

kanevry Skill

Use when encountering any bug, test failure, build break, or unexpected behavior — runs a 4-phase systematic debugging process before proposing any fix. Iron Law: NO FIXES WITHOUT ROOT CAUSE INVESTIGATION FIRST. Produces a `.orchestrator/debug/` artifact the fixer agent must reference.

kanevry Skill

Use this skill when orchestrating agentic end-to-end tests. Resolves target + profile, dispatches the right driver(s) (playwright for web today, peekaboo for macOS (issue #381)), invokes the ux-evaluator agent (opus, read-only) against driver artifacts, reconciles findings with the open issue tracker via scripts/lib/test-runner/issue-reconcile.mjs, and writes report.md + JSONL roll-up. Wraps upstream tools (no forks). Hard-gates Playwright MCP for browser drive (4× token cost vs CLI per Microsoft's own benchmark).

kanevry Skill

Creates a structured wave execution plan with role-based assignment after user alignment. Decomposes agreed tasks into waves resolved from the session mode by `scripts/session-shape.mjs`, with optimal agent assignment, dependency ordering, and inter-wave checkpoints. Activated by session-start after Q&A phase completes.

kanevry Skill

Use this skill when the user wants to identify unused, near-zero-use, or stale skills/agents/commands in the plugin surface so they can be demoted or retired. Combines agent-dispatch telemetry (start-events only) with static reference scanning, classifies every surface item into Active / Investigate / Demote / Retire, and emits a Markdown report plus JSON sidecar. NEVER auto-deletes — surfaces candidates for human decision. Quarterly cadence. <example>Context: The plugin surface has grown and the maintainer wants to prune dead weight. user: "/sunset-review" assistant: "Running the sunset walk — classifying skills, agents, and commands by usage telemetry + static refs, grouped by Retire / Demote / Investigate / Active. No item is deleted automatically; I'll surface Retire/Demote candidates for your decision." <commentary>The user wants a usage-driven prune candidate list; this skill runs the read-only walker, presents grouped verdicts, and writes a sidecar — it never deletes.</commentary></example>

kanevry Skill

Use when local resource pressure would shrink or coordinator-direct a wave, a wave plan carries heavy build/test/audit roles (test, ui, perf), or the operator says offload, remote host, or auslagern — reference for routing that wave role to a declared SSH-reachable host instead of reducing agent count

kanevry Skill

Use this skill when orchestrating documentation generation and updates within a session. Maps session scope to audience-specific docs tasks (User / Dev / Vault), dispatches the docs-writer agent with source-grounded prompts, and reports coverage gaps to session-end. Gated on `docs-orchestrator.enabled: true` in Session Config. Zero overhead when disabled.

kanevry Skill

Injects the canonical vault frontmatter schema snippet into agent prompts before any vault-write task, preventing malformed YAML frontmatter in Obsidian notes. <example>Context: wave-executor is about to dispatch a vault-mirror agent that writes learning notes under ~/Projects/vault/40-learnings/. user: "dispatch vault-write agent" assistant: "Injecting frontmatter-guard snippet into agent prompt (vault scope detected). Required fields: id, type, created, updated. Enum type: note|daily|project|person|reference|idea|learning|session." <commentary>The wave-executor pre-dispatch hook calls detectVaultTaskScope() — the fileScope contains /Projects/vault/40-learnings/ so the guard triggers and the snippet is prepended to the agent system prompt.</commentary></example>

rulebase-co Skill

Use to assess whether support conversations can be used to train, fine-tune or evaluate a model, and to build a filtered dataset if they can. Trigger for "can we train on our support data", "use transcripts to fine-tune a model", "build an eval set from real conversations", "is it ok to send our tickets to a model provider", or preparing support data for machine learning.

rulebase-co Skill

Use to find agent-side friction — tooling, permissions, dead ends, macro gaps, system hops — by shadowing real work and counting cost per contact before defaulting to training. Trigger for "agent experience audit", "why is handle time so high", "tooling friction", "agents switching systems", "shadow an agent", macro gaps, permission blocks, or "we need more training" when the stack is the problem.

rulebase-co Skill

Use to build an evaluation set and regression process for a customer-facing AI support agent, so prompt and model changes can be shipped without silently breaking answers. Trigger for "how do we test our support bot", "did the new prompt make it better", "evaluate our AI agent", building an eval set, comparing model versions, or an AI agent going to production without a test suite.

rulebase-co Skill

Use to review a support agent's draft reply before it is sent — checking factual and policy correctness first, then commitments and risk, then tone against the team's own documented principles. Trigger for "can I say this", "is this ok to send", "how should I phrase this to the customer", "is this in line with our tone of voice", an agent pasting a draft response, or asking how to explain something to a customer.

rulebase-co Skill

Use to analyse contact arrival distributions and choose staffing models that match reality — burstiness, batch dumps, abandonment censoring and when Poisson or Erlang assumptions fail. Trigger for "arrival pattern analysis", "are arrivals Poisson", "Erlang assumptions", batch email arrivals, burst traffic, abandonment bias in arrivals, staffing model choice, or when Erlang staffing misses despite a good forecast.

rulebase-co Skill

Use to measure first contact resolution and repeat contact rate defensibly, from conversation data rather than from agent-set dispositions. Trigger for "what's our FCR", "how many customers come back", "repeat contact rate", "are we resolving issues first time", one-and-done rate, or an FCR number that looks implausibly high.

rulebase-co Bundle

Use to model what a proposed SLA or response-time target would have delivered on historical data before committing to it, with correct handling of open tickets and business hours. Trigger for "what would our SLA attainment be at 2 hours", "what if we tightened the target", "should our P1 target be 2h or 4h", "% of tickets resolved within X", comparing thresholds, or setting a first-response or resolution target for the first time.

rulebase-co Skill

Use to validate an AI QA grader against human reviewers before trusting it for triage or release decisions — agreement by segment, systematic bias, and where automation can replace humans versus where it cannot. Trigger for "can we trust the AI grader", "how accurate is our automated QA", "validate the grader before rollout", "AI vs human agreement by channel or language", "is the grader too harsh on chat", or checking a grader after a model change.

rulebase-co Skill

Use to check whether ongoing QA sampling is fair to agents and still representative of the eligible population, and to fix designs that equalise by ticket instead of by person. Trigger for "is our QA sampling fair", "high-volume agents get more reviews", "sampler drift", stratified sampling by queue or team, per-agent evaluation counts, or comparing what was reviewed against what was eligible.

rulebase-co Skill

Use when QA data is being assembled for a formal employment process — a performance review, an improvement plan, a promotion case or an HR record — where the evidence standard and fairness controls are stricter than for coaching. Trigger for "document this agent's performance for HR", "build a performance file", "prepare a six-month review", "evidence for a performance improvement plan", or comparing an agent against a peer to show a gap.

rulebase-co Skill

Use to bound what an AI support agent may do to a customer's record — allowlists, irreversible actions, dry-run, audit trails, and blast-radius limits aligned with mutation-safety practice. Trigger for "what can the bot change on an account", "agent tool permissions", "AI agent safety bounds", "dry run before refund", "limit bot actions", or reviewing tool access before giving an agent write APIs.

rulebase-co Bundle

Use to measure and diagnose grader agreement in support QA — human vs human, or an AI grader vs human reviewers — separating random disagreement from one grader being systematically harsher. Trigger for "run a calibration", "how do our reviewers compare", "is the AI grading too harshly", "our QA scores have too many false positives", "why do reviewers disagree", contested or overturned evaluations, or checking a new grader or model version.

rulebase-co Skill

Use to analyse where LLM spend goes in support automation — cost per resolution, prompt bloat, retrieval size, retry loops, and the cost-vs-quality frontier from your own usage, not invented benchmarks. Trigger for "LLM cost per ticket", "why is our bot so expensive", "token usage breakdown", "reduce AI support costs", "cost vs quality tradeoff", or attributing spend to prompt, retrieval, or retries.

rulebase-co Skill

Use to measure whether an agent-assist or copilot actually helps — beyond acceptance rate — through edit distance, time and quality trade-offs, and bad-suggestion uptake. Trigger for "is our copilot working", "measure agent assist ROI", "acceptance rate is high but is it good", "did the draft suggestion help", "evaluate our support copilot", or planning an A/B rollout of assist features.

rulebase-co Skill

Use to audit retrieval and grounding as the ceiling on bot answer quality — separating missing knowledge from wrong chunks, and measuring retrieval apart from generation. Trigger for "why does the bot make things up", "RAG audit", "wrong article cited", "retrieval quality", "bot answers not in the KB", "grounding failures", or KB gaps vs search failures.

rulebase-co Skill

Use to mine resolved support conversations for missing or weak help articles, extracting reusable answers without copying PII. Trigger for "what articles should we write from tickets", conversation mining for KB, turn agent replies into help content, high-repeat contact drivers needing docs, knowledge creation from support, or closing KB gaps from real resolutions.

rulebase-co Bundle

Use to prove a conversation export is complete against the source system's own counts, before anything is built on it. Trigger for "is this export complete", "reconcile the export against Zendesk", "our export is missing tickets", "the numbers don't match the helpdesk", verifying a migration or a backfill, or signing off a dataset for analysis.

rulebase-co Skill

Use to separate customer help centre content from internal agent wiki, define what belongs where, and stop wrong internal docs from leaking into customer-facing answers. Trigger for "help centre vs internal wiki", agent documentation duplication, what should be public KB, internal SOP leaking to customers, policy vs procedure split, or grounding AI on the wrong knowledge base.

charleswiltgen Bundle

Use when implementing, testing, or evaluating ANY Apple Intelligence, on-device AI, or speech-to-text feature. Covers Foundation Models, @Generable, LanguageModelSession, Tool protocol, eval suites, model-as-judge scoring, SpeechTranscriber, CoreML.

charleswiltgen Bundle

Use when storing credentials securely, encrypting data, implementing passkeys, securing AI/agentic features against prompt injection, code signing, or managing certificates and provisioning profiles.

agentskillexchange Skill

Salesforce MCP Server

agentskillexchange Skill

ClickHouse Query Agent

agentskillexchange Skill

Notion Workspace Agent

agentskillexchange Skill

Stripe Reporting Agent

agentskillexchange Skill

Vercel Deployment Agent

agentskillexchange Skill

Devcontainer Setup Agent

agentskillexchange Skill

Elastic / Kibana MCP Server

agentskillexchange Skill

Jira Advanced Query Agent