r3f-loaders covers react three fiber asset loading with useGLTF and useLoader hooks. minimal guidance on suspense patterns, preloading strategy, or error handling in production contexts.

indexion-segment chunks text by divergence, tfidf, or punctuation for rag and embedding workflows. supports file input/output with pluggable strategies but lacks procedural depth and error handling.

brand-review evaluates marketing content against brand guidelines and style rules, with optional fallback to generic clarity and consistency checks. the skill provides structured frameworks for defining brand voice but the execution pathway is underspecified.

firecrawl-research-papers orchestrates literature review assembly by inferring topic constraints from context, searching paper corpora via firecrawl research tools, expanding via related-paper graphs, and synthesizing findings into a sourced markdown report with themes, debates, and emerging trends.

alicloud-compute-fc-agentrun covers runtime and sandbox lifecycle via aliyun openapi, including endpoint creation, credential setup, and resource querying for function compute agent workloads.

voice-isolator removes background noise and isolates speech from mixed audio via elevenlabs api. suitable for cleaning noisy recordings and preparing audio for transcription.

angular-architect provides a workflow skeleton for enterprise angular development covering standalone components, ngrx state management, and performance optimization. execution lacks specificity and edge case coverage.

web-scraping-automation covers http requests, html parsing, browser automation, and api integration across python and javascript. includes resource cleanup requirements, anti-scraping strategies, and data storage patterns without prescriptive decision logic.

gmgn-cooking covers token creation and launch on solana/bsc bonding curve launchpads via cli, with required private key setup and asynchronous polling for confirmation.

media-kit-page-generator guides creation of journalist-facing asset repositories and press pages, covering content structure, accessibility standards, and brand consistency. scopes assets like logos and visual identity while distinguishing media kits from press coverage aggregations.

skill-optimizer provides a meta-framework for diagnosing and iterating on ai skill performance via measurement loops, but defers concrete procedure to external rule files rather than self-contained guidance.

hqq-quantization enables calibration-free weight quantization across 8/4/3/2/1-bit precision levels with multiple backend optimizations, designed for rapid model compression without sample datasets.

memory-notes provides guidance on structuring markdown knowledge notes with frontmatter and semantic relations for parsing into a searchable graph, but lacks procedural depth and concrete examples for building such notes.

ln-110 coordinates project documentation generation through single-pass context gathering and type detection, delegating work to five downstream workers and aggregating results for the quality repair loop.

encore-database covers basic sql database initialization and query method introduction without substantive guidance on migrations, error handling, or integration patterns.

deepchem covers molecular property prediction and featurization for chemistry and biology workflows, with access to pretrained models and benchmark datasets, but lacks procedural depth for end-to-end implementation.

readme generates comprehensive project documentation by outlining structure and purpose but lacks concrete step-by-step procedure, decision logic, and explicit failure handling.

passkit-wallet covers apple pay button display, payment request construction, authorization handling, and wallet pass creation for ios. lacks procedural depth and error recovery guidance.

arize-compliance-audit scans ai apps against eu ai act, gdpr, nist ai rmf and other frameworks, producing tailored checklists and remediation guidance. outputs are advisory only and require legal review.

avoid-ai-writing audits prose for machine-generation markers and offers rewrites. cautious framing acknowledges false-positive risks and the gap between statistical signals and evidentiary proof.

forge extracts candidate learnings from session transcripts and queues them for filtering by a promotion ratchet. runs via sessionend hook to populate a learnings registry without requiring manual intervention.

structured-autonomy-implement executes a provided implementation plan sequentially without deviation, validating progress against explicit steps and halting on stop signals.

quick-stats generates inline backtests for indian equities using ema crossover strategies, fetches ohlc data from openalgo with yfinance fallback, applies delivery fees, and outputs compact performance metrics with plotly equity curves.

competitor-profiling scrapes competitor websites and aggregates seo metrics, review site data, and pricing information into structured markdown profiles. supports variable depth research and multi-competitor parallelization.

gemini-imagegen wraps google's gemini api for text-to-image generation and editing with configurable resolution and aspect ratios, but lacks procedural clarity and failure handling.

system-design guides solo developers through architecture decisions by surfacing requirements clarity and trade-offs, but lacks complete procedures for translating validated needs into concrete component designs.

critical-code-reviewer applies adversarial scrutiny to code submissions, operating from first-principles distrust and demanding concrete evidence of quality. incomplete specification undermines deployment clarity.

argent-test-ui-flow orchestrates cross-platform mobile ui testing via argent mcp tools, delegating platform setup to ios or android-specific skills and using device selection heuristics to route interactions.

launchdarkly-guarded-rollout guides configuration of progressive feature releases with staged traffic increases and metric monitoring, but lacks actionable step-by-step procedure and explicit rollback criteria.

code-review covers receiving technical feedback critically, requesting systematic reviews via subagent, and enforcing verification gates before accepting completion claims. emphasizes evidence-based practices over performative agreement.

react-native-testing covers rntl v13 and v14 test authoring with version-aware api handling and query selector prioritization. lacks complete procedures for common test scenarios.

model-usage parses CodexBar cost logs to extract per-model usage summaries for Codex or Claude, supporting current or aggregate breakdowns via Python CLI.

pricing-strategy guides users through designing iterative pricing approaches using frameworks distilled from 46 product leaders, emphasizing willingness-to-pay research and common transition points across business model stages.

cs-note appends project micro-knowledge to .codestable/attention.md via fixed sections, capturing short-lived-but-essential facts like build steps, service startup requirements, and command aliases that ai needs on every launch.

ce-test-browser orchestrates end-to-end browser tests against pages changed in a pr or branch, routing all automation through the agent-browser cli. requires local dev server and git repo setup.

writing-job-descriptions guides users through a framework centered on 12-month business outcomes, identifying candidate spike skills, and using intentional language to filter cultural fit, drawing from product leader practices.

firecrawl-monitor detects website content changes and delivers notifications via webhook or email, with per-field diffs and change classification (same, new, changed, removed, error).

backend-to-frontend-handoff-docs guides api documentation creation for frontend consumption, covering context transfer, endpoint specs, and validation rules. trigger coverage is narrow and language is often generic rather than prescriptive about what goes in each section.

cloudflare covers the entire platform via decision trees that route to compute, storage, ai, networking, and security products. authentication and deployment verification are required before use.

conversational workflow for crafting system prompts within the sanity agent context mcp framework. assumes user control over system prompt configuration and defaults to instructions field as primary behavior lever.

author-contributions identifies all files modified by a specified git author across a feature branch versus upstream, returning a markdown table with change status and line deltas.

requesthunt aggregates user feedback from reddit, x, github, youtube, linkedin, and amazon to generate demand research reports via structured scraping and analysis with rate-limited api access.

chat-with-pdf enables question-answering, summarization, and extraction from pdf documents with citations and confidence levels. handles multi-document comparison and common layout challenges but lacks detail on actual implementation or api integration.

use-cases-page-generator helps teams create marketing pages that connect specific product features to customer problems, organized by scenario and persona rather than industry taxonomy.

hyperframes-media provides three local cli commands for media preprocessing: text-to-speech with 54 multilingual voices, audio transcription with word-level timestamps, and video background removal for transparent overlays. no api keys required.

agent-manager-skill orchestrates multiple cli agents across tmux sessions with task assignment, real-time monitoring, and cron scheduling for parallel local workflows.

safe-action-client covers initialization and basic action scaffolding with zod validation in next.js server environments, but stops short of handling errors, configuration patterns, or failure recovery.