prompt-fidelity
Self-check skill: decompose a request into verified vs inferred constraints, compute a fidelity score, and report which parts of an answer are auditable vs judgment calls.
Lot de plugins · par James Barney
Aperçu
prompt-fidelity est un lot de plugins pour les outils de code IA publié par James Barney. Places de marché que nous lisons et qui le référencent : 1 (Community GitHub marketplaces). Nous l'avons vu pour la première fois le 2026-10-14 et n'avons enregistré aucun changement depuis. Selon notre passage sur les places de marché du 2026-10-14.
Ce qu'il demande
- S'installe comme lot de plugins
Aucune des places de marché que nous lisons ne publie ses outils ni les réglages qu'il demande.
Ce qu'il déclare, demande et exécute, en détail
N’installez que des plugins de développeurs en qui vous avez confiance. Nous reprenons ce que publient les places de marché et ne testons pas les plugins.
Fonctionne avec
- Claude Code
D'après les marketplaces qui le listent et le transport MCP qu'il déclare, tels que chaque client les documente. Nous ne testons pas les plugins.
Connecter
Commandes et configurations construites uniquement à partir de ce que déclarent ses places de marché. Nous ne les exécutons pas pour vous et ne testons pas les plugins.
En plugin Claude Code
Construit à partir de ce que déclare Community GitHub marketplaces.
/plugin marketplace add barneyjm/prompt-fidelity
/plugin install prompt-fidelity@prompt-fidelityOù il est publié
| Place de marché | Statut | Version | Chiffres | Vérifié | Vu pour la première fois |
|---|---|---|---|---|---|
| Place de marché communautaire barneyjm/prompt-fidelity (Ouvrir la source sur GitHub) | Référencé | 0.1.0 | Aucun chiffre publié | Non marqué |
Chaque chiffre est celui de la place de marché et n'est jamais additionné entre places de marché. Une place de marché qui ne publie pas de chiffre n'en affiche aucun : pas de nombre ne veut pas dire zéro.
Listé dans une place de marché communautaire
| Place de marché | Dépôt | Étoiles (GitHub) |
|---|---|---|
| prompt-fidelity | barneyjm/prompt-fidelity | 1 |
Quelles places de marché nous lisons et comment nous les vérifions
Ce que contient le lot
Tel que publié sur Community GitHub marketplaces.
Skills: 1
prompt-fidelity
Ce qui a changé
- Est devenu un produit que nous suivons
- Community GitHub marketplaces: apparu
Plugins similaires
Produits dont les noms, descriptions, outils et skills sont les plus proches de celui-ci par le sens, sur toutes les places de marché que nous lisons. La proximité n'est classée qu'à l'intérieur de cette liste.
-
Three-layer verification pipeline for AI output. Extracts claims, finds sources, and flags hallucination risks so humans can verify before acting.
-
HallucC — Hallucination & Safety Guardian
Fact-check LLM outputs with cited sources. API key: aihcc.cloud/keys.
-
Hallucination validator and factual grounding checker for LLM outputs.
-
Spectral — verified-knowledge oracle
Cross-check a factual claim against a verified knowledge graph before you assert it. Never guesses.
-
Independent fact-checking for AI answers: a verdict and sources for every claim.
-
Pay-per-call AI evaluation MCP server. Score LLM outputs against benchmark rubrics via Workers AI.
-
Ask other AI agents, get a second opinion, and publish or challenge knowledge claims on ANP2's public signed log.
-
Local-first MCP server for LLM output verification — risk signals, injection and PII checks.
-
Grounding/hallucination guard: verify a claim against evidence (Supported/Refuted/NEE).
-
Score two AI outputs, get exact failure reasons, and issue a $0.01 USDC receipt via x402.
Plugins similaires par place de marché
La même question posée à chaque place de marché séparément : y a-t-il là un équivalent de ce plugin ?
Référencé sur Claude Code: claude-plugins-community (5)
-
Fortytwo Prime: collective multi-agent inference for high-stakes questions. ALWAYS use when the user says "Ask Fortytwo", "Ask Prime", "Ask Fortytwo Prime", or calls…
-
Context engineering plugin for Claude Code — real-time prompt scoring, RLM adaptive weights, and token optimization. Scores every tool use across 5 dimensions (token utilization…
-
Break through difficult problems by gathering diverse AI perspectives in parallel. Consensus queries multiple AI providers concurrently — GPT-5.2, Gemini 3.1 Pro, and DeepSeek…
-
Skills for working with Langfuse, the open-source LLM engineering platform for tracing, prompt management, and evaluation.
-
AI agent evaluation sandbox for Claude Code. Holdout scenario testing with Doer/Judge/Adversary/Observer roles, probabilistic satisfaction scoring, and append-only JSONL audit…
Référencé sur Claude Code: claude-plugins-official (5)
-
Skills for adding DeepEval evaluations, tracing, datasets, Confident AI reports, and iterative improvement loops to AI applications.
-
Skills for working with Langfuse, the open-source LLM engineering platform for tracing, prompt management, and evaluation.
-
Output.ai workflow development toolkit for Claude Code. Adds 5 specialist agents (planner, builder, debugger, prompt writer, quality reviewer), 40+ slash-command skills covering…
-
Generate an explorable HTML report of Claude Code session usage — tokens, cache efficiency, subagents, skills, and the most expensive prompts — from local ~/.claude/projects…
-
You.com agent skills for web search, research with citations, and content extraction. Guided integrations for Vercel AI SDK, Claude Agent SDK, OpenAI Agents SDK, crewAI…
Référencé sur Claude: knowledge-work-plugins (5)
-
Skills for working with Langfuse, the open-source LLM engineering platform for tracing, prompt management, and evaluation.
-
Exa AI web search, deep research, and content extraction. Provides MCP tools and research skills for comprehensive web search, people discovery, company research, academic papers…
-
AI-first project auditor and re-engineer based on the 9 design principles and 7 design patterns from the TechWolf AI-First Bootcamp
-
Build multimodal AI applications with Pixeltable -- tables, computed columns, embedding search, UDFs, tool-calling agents, and 25+ AI provider integrations.
-
Connects Claude Code to the Synthflow AI voice-agent platform through its hosted MCP server, with skills for reviewing calls and auditing agent prompts, plus a docs-search…
Référencé sur Codex: openai-curated (openai/plugins) (5)
-
Evaluate Codex skills and plugins from chat with a beginner-friendly start command, local-first reports, token budget explanations, and guided benchmarking.
-
Answer product and business questions with data
-
Build with OpenAI APIs, Agents SDK, and ChatGPT Apps, and create and save OpenAI API keys from Codex.
-
PostHog gives your AI agent direct access to your product analytics, feature flags, experiments, error tracking, surveys, logs, and LLM analytics. Ask questions about your data…
-
Notion workflows for implementation planning, research synthesis, meeting preparation, and knowledge capture.
Référencé sur Community GitHub marketplaces (5)
-
Ask other AI agents, get a second opinion, and publish or challenge knowledge claims on ANP2's public signed log.
-
Skills for adding DeepEval evaluations, tracing, datasets, Confident AI reports, and iterative improvement loops to AI applications.
-
A runtime-portable 5-agent quality gate that adjudicates text artifacts before they ship. Five role-conditioned LLM deliberators run a 2-round async protocol with cross-read…
-
Multi-model LLM Council and parallel AI window for Claude Code. Fork any model, fold results back.
-
Screen prompts, skills and MCP tools with TypeSafe judgments and configurable reasoning models.
Référencé sur Copilot CLI: github/awesome-copilot (5)
-
Three-layer verification pipeline for AI output. Extracts claims, finds sources, and flags hallucination risks so humans can verify before acting.
-
A runtime-portable 5-agent quality gate that adjudicates text artifacts before they ship. Five role-conditioned LLM deliberators run a 2-round async protocol with cross-read…
-
Goal-driven task orchestration with independent verification. Interviews the user to define a clear goal, then loops between a Builder subagent (does the work) and an Inspector…
-
Run a lightweight, role-separated AI development team with flexible tools, developer-selected models, proportionate planning, and optional QA.
-
Visualizes live session AI-credit usage as a Pac-Man board with pellets, ghosts, fruit milestones, and game-over limits.
Référencé sur Copilot CLI: github/copilot-plugins (5)
-
Toolkit for building Microsoft 365 Copilot declarative agents — scaffolding, JSON manifest development, and capability configuration.
-
workiq
Query Microsoft 365 data with natural language — emails, meetings, documents, Teams messages, and more.
-
workiq-productivity
WorkIQ productivity skills powered by the local WorkIQ CLI — read-only insights across email, meetings, calendar, Teams, SharePoint, projects, and people.
-
Advanced Security plugin for GitHub Copilot.
-
Build Power Apps Canvas Apps using the Canvas Authoring MCP server.
Référencé sur Cursor: cursor/plugins (5)
-
Search company knowledge and draft verified answers.
-
Iterative self-referential AI loops using the Ralph Wiggum technique.
-
Track AI visibility, sentiment, and citations.
-
Consult a stronger model before major decisions, when stuck, and before declaring done.
-
Incremental transcript-driven memory updates for AGENTS.md using high-signal bullet points only.
Référencé sur Docker MCP Catalog (5)
-
Read and score Opik traces, experiments, datasets and prompts from your AI assistant.
-
Real-time web search, reasoning, and research through Perplexity's API
-
Production-ready RAG service to search and retrieve data from your documents.
-
Sequential Thinking (Reference)
Dynamic and reflective problem-solving through thought sequences.
-
Exa AI web search, deep research, and content extraction. Provides MCP tools and research skills for comprehensive web search, people discovery, company research, academic papers…
Référencé sur Official MCP Registry (5)
-
HallucC — Hallucination & Safety Guardian
Fact-check LLM outputs with cited sources. API key: aihcc.cloud/keys.
-
Hallucination validator and factual grounding checker for LLM outputs.
-
Spectral — verified-knowledge oracle
Cross-check a factual claim against a verified knowledge graph before you assert it. Never guesses.
-
Independent fact-checking for AI answers: a verdict and sources for every claim.
-
Pay-per-call AI evaluation MCP server. Score LLM outputs against benchmark rubrics via Workers AI.
Trouvés par le sens, et non par un code ou un dépôt commun : les plugins similaires restent des produits distincts, et ceci n'est ni un avis ni une recommandation.
C'est votre plugin ?
Ajoutez-le à vos projets et prouvez que vous le maintenez : un fichier dans son dépôt, un fichier sur votre domaine ou un enregistrement DNS. Vous recevez des alertes quand une marketplace le publie, le retire, change sa marque vérifiée ou ce qu'il demande, et son endpoint distant est surveillé comme un service web.
Données issues de Community GitHub marketplaces. Les chiffres proviennent de nos propres passages réguliers sur les places de marché, dernière lecture le . Nous ne publions aucune coordonnée des éditeurs de plugins.