「sam」の検索結果

Improve agent reliability over time — diagnose why agents fail and fix the setup. Triggers on: agent keeps failing, same mistake again, agent not improving, make agent smarter, agent quality plateau, agents ignore skills, agent skips tests, fix agent behavior, agent unreliable, improve agent setup, self-improving harness, agents worse over time, tune agent instructions, agent going in circles, agent ignores AGENTS.md, repeated agent errors. Requires harness v0 and eval harness. AUTO-ROUTED from harness-engineering on symptoms. Not first setup — harness-generation first.

日本語の概要は準備中です。原文の説明を表示しています。

Bootstrap full agent infrastructure for an existing, already-coded project without modifying a single line of source code. Surveys the repo, infers conventions from manifests, README, git history, and source samples, then generates the missing AGENTS.md, docs/architecture.md, docs/product-soul.md, ADR backfill, and seeded docs/memory/ files. Asks the user only about gaps the repo cannot reveal. Load when the user asks to "retroactive project setup", "backfill agent infrastructure", "bootstrap agents for this existing repo", "onboard agents to a legacy codebase", "set up agents without touching code", or "fill in missing agent context for this project".

日本語の概要は準備中です。原文の説明を表示しています。

Orchestrator for agent harness work — the setup that makes AI agents follow project rules and improve when they fail. FIRES PROACTIVELY when agents misbehave, repeat mistakes, ignore instructions, skip skills, or when AGENTS.md exists but docs/harness/manifest.json is missing. Also triggers on: harness engineering, agent scaffold, agent keeps failing, agent not following instructions, make agents reliable, agents going off rails, agent forgot context, improve agent setup, self-improving agents, agents keep making mistakes, why is my agent bad, agent quality, agent setup broken, agents ignore skills, same mistake again, fix agent behavior, tune agent instructions, set up agent infrastructure, after project setup agents still bad. Routes bootstrap vs evolution. Not multi-agent topology — agent-builder.

日本語の概要は準備中です。原文の説明を表示しています。

Write a rigorous, decision-grade experiment spec — falsifiable hypothesis, primary metric, guardrails, randomisation unit, exposure definition, method (A/B, holdout, switchback, quasi-experiment, MAB), MDE/duration plan, peek policy, validity threats, and pre-committed decision rule. Platform-agnostic. Load when the user has a candidate experiment and needs to spec it before launch, or says "spec this experiment", "write the test plan", "design this A/B test", "what's the hypothesis", "how big a sample do we need", "how long should we run this", "define the metrics for this test", or when the experimentation orchestrator routes here.

日本語の概要は準備中です。原文の説明を表示しています。

Cleans up local git branches marked as [gone] (deleted on the remote but still present locally), including removing any associated worktrees. Use when the user says "prune branches", "clean up gone branches", "/prune-branches", or asks to delete stale local branches whose remotes are gone.

日本語の概要は準備中です。原文の説明を表示しています。

Audit UI, docs, or prose for AI-generated tells in design and copy, then fix approved findings. Use when the user says de-slop, slop check, "looks AI-generated", or wants copy humanized.

日本語の概要は準備中です。原文の説明を表示しています。

Create atomic git commits with terse, exact Conventional Commits messages. Cuts noise, preserves intent. Subject ≤50 chars. Body only when "why" isn't obvious. Use when the user says "commit", "commit this", "/commit", or asks to stage and commit changes.

日本語の概要は準備中です。原文の説明を表示しています。

Generate a single unified changelog from git history that works for everyone in the company (managers and non-tech colleagues as well as developers). Outputs Slack mrkdwn so it can be pasted straight into Slack. Defaults to Slovak (English only on explicit request) and to comparing the release branch (main) against develop to show unreleased changes; also supports date ranges like "today", "yesterday", "last week", or an explicit git range. User-invoked only via /changelog.

日本語の概要は準備中です。原文の説明を表示しています。

Monitor a pull request through review and CI until it's green: verify every bot claim against the code, fix what's real, push back on what's wrong, rerun flaky checks. Use when the user asks to babysit, monitor, watch, or shepherd a PR, says "get this PR green" or "handle the review comments". Accepts an optional PR number/URL; defaults to the current branch's open PR.

日本語の概要は準備中です。原文の説明を表示しています。

Sync the live ~/.claude global config (settings.json, CLAUDE.md, statusline.ts, own skills, skills.sh lock file) into this git backup repo, then commit. Allowlist-based so credentials, history, and project transcripts are never copied. Use when the user says "sync my claude config", "back up .claude", "update the .claude repo", or runs "/sync-claude".

日本語の概要は準備中です。原文の説明を表示しています。

Implement several variations of a design directly inside the running app, on the real page/route, with a live switcher to toggle between them in the browser, then recommend one and fold the winner into the real code. Variants are real implementations judged against real data, layout, and density (not an isolated mockup). Use this skill whenever the user wants to try design options live in the app, prototype changes directly on a real page, add a variant switcher, compare layouts in-context, asks to "build a few versions of this page", "try variants on the real screen", "switch between designs in the app", "implement this with a toggle", or says "real-prototype this". For a safe standalone .html copy that never touches the app, use the prototype skill instead.

日本語の概要は準備中です。原文の説明を表示しています。

Verify a code change end-to-end in the current worktree: identify what changed, run the relevant tests, get the app running, drive a browser over the affected areas, and report a verdict with screenshots. Generic across project types (Laravel/Herd, Node/bun, frontend, CLI, library). Use when the user wants to verify/QA/smoke-test a change, check that a PR or branch works, validate uncommitted work, or says "/flow", "verify this change", "does this PR work", "test and click through this". Optional target: PR number/URL, branch/ref, or worktree path; default is the current working changes.

日本語の概要は準備中です。原文の説明を表示しています。

Type-safe ORM for Cloudflare D1 using Drizzle. Use when building D1 schemas, writing type-safe queries, generating migrations with Drizzle Kit, defining relations, using db.batch, or hitting D1_ERROR, BEGIN TRANSACTION failures, foreign key constraints, env.DB undefined, or schema inference issues.

日本語の概要は準備中です。原文の説明を表示しています。

Create branch, atomic commits, push, and open a pull request with terse, exact messages. Conventional branch + commit format. PR title ≤70 chars, body says "why", not "what". Use when the user says "create pr", "create pullrequest", "create pull request", "make pr", "open pr", "open pullrequest", "open pull request", "submit pr", "raise pr", "/create-pr", or asks to ship/send changes as a pull request (PR/pullrequest/pull-request, any spelling).

日本語の概要は準備中です。原文の説明を表示しています。

Create atomic git commits with terse Conventional Commits messages, then push to remote. Subject ≤50 chars, body only when "why" isn't obvious. Never force pushes. Use when the user says "commit and push", "commit & push", "/commit-push", or asks to stage, commit, and push in one flow.

日本語の概要は準備中です。原文の説明を表示しています。

Apple's approach to interface design and fluid, physical motion, translated for the web. Use when building or reviewing gesture-driven UI, spring animations, drag/swipe/sheet interactions, momentum and interruptible transitions, translucent materials and depth, typography (optical sizing, tracking, leading), reduced-motion, or the design foundations (feedback, spatial consistency, restraint) behind Apple-style interfaces.

日本語の概要は準備中です。原文の説明を表示しています。

Senior embedded software engineer specializing in firmware and driver development for ARM Cortex-M microcontrollers (Teensy, STM32, nRF52, SAMD).

日本語の概要は準備中です。原文の説明を表示しています。

Convert human-written CLAUDE.md into AI-native structured-label format. Battle-tested across 4 models. Same rules, fewer tokens, higher compliance.

日本語の概要は準備中です。原文の説明を表示しています。

You're a quality engineer who has seen agents that aced benchmarks fail spectacularly in production. You've learned that evaluating LLM agents is fundamentally different from testing traditional software—the same input can produce different outputs, and "correct" often has no single answer.

日本語の概要は準備中です。原文の説明を表示しています。

Assembles microbial-community sequencing into metagenome-assembled genomes (MAGs) with metaFlye (ONT), metaSPAdes/MEGAHIT (Illumina), and hifiasm-meta/metaMDBG (PacBio HiFi), then recovers genomes via multi-binner consolidation (MetaBAT2, MaxBin2, CONCOCT, SemiBin2, VAMB -> DAS_Tool) and QCs them against MIMAG with CheckM2, GUNC, and GTDB-Tk. Covers why a metagenome is not a genome (uneven coverage, micro-diversity, strain collapse to consensus), differential-coverage binning, co-assembly vs per-sample, the rRNA-operon collapse that fails short-read MAGs, and strain resolution with inStrain. Use when reconstructing genomes from a microbiome, soil, ocean, or gut community, recovering MAGs, or resolving strain-level variation.

日本語の概要は準備中です。原文の説明を表示しています。

Assembles haplotype-resolved diploid and telomere-to-telomere (T2T) genomes from PacBio HiFi reads with hifiasm (HiFi-only, Hi-C, or trio phasing) and verkko (HiFi + ultralong ONT for T2T), extracting contigs from GFA and routing phasing QC to k-mer/trio metrics. Covers why a primary assembly is a haplotype mosaic that exists in no cell, partial-vs-full phasing (the .bp. vs .dip. filename convention), the purge-default trap on inbred samples, the --hom-cov coverage-estimate alarm, and verkko-vs-hifiasm for T2T. Use when assembling a diploid eukaryote from HiFi, phasing haplotypes with parents (trio) or Hi-C, deciding whether to chase T2T, or diagnosing switch errors invisible to N50/BUSCO/QV.

日本語の概要は準備中です。原文の説明を表示しています。

Quality control for IMC/MIBI data across pixel, channel, image, slide, and batch levels, covering Poisson-count SNR (cell-level Gaussian-mixture and empty-channel comparison), spillover-matrix QC (the three physical sources), drift and the missing EQ-bead analog, acquisition artifacts, and sample-of-origin batch effects. Use when deciding whether to keep or drop a channel, ROI, or slide, distinguishing a dim antibody from a failed one, reading a spillover matrix, or diagnosing batch-driven clustering before analysis.

日本語の概要は準備中です。原文の説明を表示しています。

Identifies differential m6A methylation between conditions from MeRIP-seq paired IP/input data using exomePeak2 with `bam_ip` + `bam_input` (control arm) and `bam_treated_ip` + `bam_treated_input` (treatment arm) for integrated GC-bias-aware differential calling (Liu 2022 *NAR Genom Bioinform* 4:lqac046), QNB beta-binomial test (Liu 2017 *BMC Bioinformatics* 18:387), MeTDiff HMM-based differential bundled with MeTPeak, RADAR (Zhang 2019 *Genome Biol* 20:294) with its `filterBins -> diffIP -> reportResult` workflow, and the defensible paired-symmetric case of edgeR / DESeq2 on featureCounts-on-peaks matrices when batch / lot covariates need explicit fixed-effect handling (exomePeak2's top-level API does NOT accept arbitrary covariates). Covers paired vs unpaired vs interaction designs, batch confounding (antibody lot, RNA prep, sequencing run) and the per-lot meta-analysis strategy when exomePeak2 is the primary caller, the stoichiometry-vs-expression-vs-IP-efficiency confound that all MeRIP differential methods inherit, normalisation choice (size factor on IP, on input, on per-sample IP/input ratio), the McIntyre 2020 reproducibility caveat, effect-size filtering as a guardrail against under-powered N=2 designs, and orthogonal-validation routes for absolute stoichiometry (GLORI / SAC-seq / m6Anet `mod_ratio`). Use when comparing m6A levels across two or more conditions, choosing between exomePeak2 / QNB / RADAR / MeTDiff for a given design, handling batch confounding when exomePeak2's API is too rigid, normalising against input properly, distinguishing real hyper- / hypo-methylation from expression-level shifts, applying effect-size thresholds, interpreting volcano plots of differential peaks, or planning a follow-up orthogonal stoichiometry validation.

日本語の概要は準備中です。原文の説明を表示しています。

Aligns and QCs methylated-RNA-immunoprecipitation (MeRIP / m6A-seq) IP and input libraries using STAR or HISAT2 splice-aware mapping, samtools sort/index, IP/input matched-pair tracking, antibody-lot metadata recording, replicate concordance via deepTools multiBamSummary + plotCorrelation, IP enrichment QC via plotFingerprint and per-transcript IP/input ratio distributions, library-complexity saturation curves via PreSeq c_curve / lc_extrap, and the explicit do-NOT-deduplicate convention for standard non-UMI MeRIP. Use when preparing paired IP and input BAM files for exomePeak2 / MeTPeak / MACS3 peak calling, evaluating MeRIP replicate concordance and IP enrichment, deciding whether to deduplicate (standard MeRIP typically NOT — see Tips), choosing genome-vs-transcriptome alignment for downstream peak vs m6Anet workflows, recording antibody clone and lot metadata for downstream cross-batch reconciliation, detecting failed IPs via saturation curves and IP/input distribution shape, or generating IP-over-Input bigWig tracks for downstream visualisation.

日本語の概要は準備中です。原文の説明を表示しています。