feat(plugins): DSPy.ts Claude Code / Codex plugin marketplace#13
Open
ruvnet wants to merge 11 commits into
Open
feat(plugins): DSPy.ts Claude Code / Codex plugin marketplace#13ruvnet wants to merge 11 commits into
ruvnet wants to merge 11 commits into
Conversation
…anifest + ADR-0001 Bootstrap a Claude Code / Codex plugin marketplace from this repo: - .claude-plugin/marketplace.json — lists 10 plugins (dspy-core available; optimize/rag/react/observability/evolution + 4 vertical appliances planned) - plugins/dspy-core/ — first plugin: commands (dspy-new, dspy-compile, dspy-eval), agent (dspy-architect), skills (signature-design, metric-design), and a bundled MCP server scaffold (dspy_scaffold/dspy_compile/dspy_eval/dspy_retrieve tools + dspy:// resources) wired via plugin.json mcpServers - docs/adr/ADR-0001-claude-code-plugin-marketplace.md — the decision + roadmap Remaining plugins build one per iteration (scaffold → validate → commit), tracked in a GitHub issue, finishing with this PR. Co-Authored-By: RuFlo <ruv@ruv.net>
…rapFewShot dynamic demos Second plugin in the marketplace. Commands /dspy-mipro, /dspy-gepa, /dspy-bootstrap (thin orchestrators over MIPROv2/GEPA/BootstrapFewShot + AgentDBClient + CompilationTracer); optimizer-engineer agent; optimizer-selection + experience-replay skills; bundled MCP server scaffold (dspy_mipro / dspy_gepa / dspy_bootstrap / dspy_replay_status tools + dspy:// resources) wired via plugin.json mcpServers; AGENTS.md for Codex. Marketplace status flipped to available. Tracks #12. Co-Authored-By: RuFlo <ruv@ruv.net>
…Thought, grounded + cited Third plugin. Commands /dspy-index (chunk+embed+store an AgentDB corpus, HNSW/RaBitQ/tiers), /dspy-rag-new (scaffold a Retrieve→ChainOfThought Pipeline + grounding metric), /dspy-retrieve (debug top-k + MMR for a query). rag-architect agent; chunking-strategy + grounding-and-citations skills; bundled MCP server scaffold (dspy_index / dspy_retrieve / dspy_rag_run / dspy_corpus_status tools + dspy:// resources) via plugin.json mcpServers; AGENTS.md for Codex. Marketplace status flipped to available. Tracks #12. Co-Authored-By: RuFlo <ruv@ruv.net>
…pisodes, skill promotion) Fourth plugin. Commands /dspy-react-new (scaffold a ReAct agent + tool registry [+ reflexion on AgentDB]), /dspy-react-run (run on a task, full thought/action/observation trace + what got recorded), /dspy-reflexion (inspect lessons / promoted skills / episodes). react-engineer agent; tool-design + reflexion-loop skills; bundled MCP server scaffold (dspy_react_new / dspy_react_run / dspy_reflexion_recall / dspy_reflexion_status tools + dspy:// resources) via plugin.json mcpServers; AGENTS.md for Codex. Marketplace status flipped to available. Tracks #12. Co-Authored-By: RuFlo <ruv@ruv.net>
… MLflow) + CachingLM Fifth plugin. Commands /dspy-trace (compile with a CompilationTracer → run summary + causal chain), /dspy-runs (list/inspect compile runs in an AgentDB trace store), /dspy-cache (wrap an LM with CachingLM — fuzzy AgentDB vector cache — + hit stats). observability-engineer agent; compile-tracing + llm-caching skills; bundled MCP server scaffold (dspy_trace / dspy_runs / dspy_run_chain / dspy_cache_stats tools + dspy:// resources) via plugin.json mcpServers; AGENTS.md for Codex. Marketplace status flipped to available. Tracks #12. Co-Authored-By: RuFlo <ruv@ruv.net>
…olution against a benchmark Sixth plugin (exotic tier). Commands /dspy-evolve (N-generation GEPA evolution vs a held-out benchmark — reflect on weak cases, mutate instructions, persistent AgentDB Pareto frontier with warm-start, optional structural exploration: module swap / insert Retrieve), /dspy-benchmark (score raw/optimized/evolved variants over a held-out set, record results), /dspy-evolution-status (learning curve / reflections / lineage). evolution-coordinator agent; benchmark-design + self-evolution-loop skills; bundled MCP server scaffold (dspy_evolve / dspy_benchmark_run / dspy_evolution_status / dspy_frontier_inspect tools + dspy:// resources) via plugin.json mcpServers; AGENTS.md for Codex. Marketplace status flipped to available. Tracks #12. Co-Authored-By: RuFlo <ruv@ruv.net>
…istant (Retrieve→CoT, tuned) Seventh plugin (vertical appliance). Ships a ready-to-run DSPy.ts program template (RetrieveModule MMR over an AgentDB KB → ChainOfThought answer with citations + a quality metric: helpfulness + groundedness + honest 'I don't know') plus commands /support-bot-init (scaffold into the repo + create the KB), /support-bot-ask (retrieve + answer + cite), /support-bot-tune (MIPROv2 against a Q/A set + the metric, with replay + tracer). support-bot-builder agent; support-kb-curation + support-answer-quality skills; bundled MCP server scaffold (support_bot_init / support_bot_ask / support_bot_tune / support_bot_status tools + dspy:// resources) via plugin.json mcpServers; AGENTS.md for Codex. Marketplace status flipped to available. Tracks #12. Co-Authored-By: RuFlo <ruv@ruv.net>
…CoT review, GEPA-tuned Eighth plugin (vertical appliance). Ships a ready-to-run DSPy.ts program template (RetrieveModule pulls repo context: conventions/ADRs/related code/prior reviews → ChainOfThought structured review: severity-tagged located findings + suggested fixes → an actionability metric: coverage − noise, specificity, calibration, no false alarms on clean PRs) plus commands /code-review-init (scaffold + build the context index), /code-review-run (review a diff/file), /code-review-tune (GEPA/MIPROv2 vs labelled reviews + the metric, frontier + cache). code-review-builder agent; review-context-indexing + review-actionability-metric skills; bundled MCP server scaffold (code_review_init / code_review_run / code_review_tune / code_review_status tools + dspy:// resources) via plugin.json mcpServers; AGENTS.md for Codex. Marketplace status flipped to available. Tracks #12. Co-Authored-By: RuFlo <ruv@ruv.net>
…/note)+reflexion → cited synthesis Ninth plugin (vertical appliance). Ships a ready-to-run DSPy.ts program template (a ReAct agent over a search/fetch/note tool registry with ReActReflexion → a ChainOfThought synthesizer that writes a grounded, cited answer with named gaps → a groundedness/coverage metric: cited evidence must be gathered, covers the sub-topics, calibrated about gaps, honest 'insufficient evidence') plus commands /research-init (scaffold + create the reflexion store), /research-ask (run a query, full ReAct + synthesis trace), /research-tune (MIPROv2 vs a graded set + the metric, target synthesizer|gatherer|both, replay + tracer). research-assistant-builder agent; research-tool-registry + synthesis-and-grounding skills; bundled MCP server scaffold (research_init / research_ask / research_tune / research_status tools + dspy:// resources) via plugin.json mcpServers; AGENTS.md for Codex. Marketplace status flipped to available. Tracks #12. Co-Authored-By: RuFlo <ruv@ruv.net>
…eline + BootstrapFewShot from a labeled CSV Tenth (final) plugin (vertical appliance). Ships a ready-to-run DSPy.ts program template (typed PredictModule stages clean → classify → extract composed in a Pipeline, a CSV/JSONL batch loader with per-row error capture, a graded row metric — exact label=1, off-vocab capped, real-class↔other confusion penalised — and a BootstrapFewShot tuning helper with optional input-conditioned dynamic demos) plus commands /data-pipeline-init (scaffold + set labels/stages), /data-pipeline-run (process a batch → output CSV + summary), /data-pipeline-tune (BootstrapFewShot the classify stage from labels, held-out accuracy + confusion matrix). data-pipeline-builder agent; typed-pipeline-design + batch-classification-metrics skills; bundled MCP server scaffold (data_pipeline_init / data_pipeline_run / data_pipeline_tune / data_pipeline_eval tools + dspy:// resources) via plugin.json mcpServers; AGENTS.md for Codex. Marketplace status flipped to available — 10/10. Tracks #12. Co-Authored-By: RuFlo <ruv@ruv.net>
…10 plugins) Co-Authored-By: RuFlo <ruv@ruv.net>
ruvnet
marked this pull request as ready for review
May 11, 2026 21:01
10 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds a Claude Code / Codex plugin marketplace published from this repo. See
docs/adr/ADR-0001-claude-code-plugin-marketplace.md. Tracking: #12.Landed in this PR (so far)
.claude-plugin/marketplace.json— lists 10 plugins (1 available, 9 planned, flipped toavailableas each lands)plugins/dspy-core/— first plugin:/dspy-new(scaffold a Signature + module),/dspy-compile(run BootstrapFewShot / MIPROv2 / GEPA),/dspy-eval(score over a dataset)dspy-architect(picks module, writes Signature, designs metric, chooses optimizer)signature-design,metric-designmcp/server.js→dspy_scaffold/dspy_compile/dspy_eval/dspy_retrievetools +dspy://docs/api,dspy://examples, design-guide resources), wired viaplugin.json'smcpServers— Codex picks up the same serverdocs/adr/ADR-0001-claude-code-plugin-marketplace.mdStill to come (one plugin per iteration — see #12 checklist)
dspy-optimize,dspy-rag,dspy-react,dspy-observability,dspy-evolution(exotic), then vertical appliancesdspy-appliance-support-bot,dspy-appliance-code-review,dspy-appliance-research-assistant,dspy-appliance-data-pipeline. Marked ready for review once all are in. Closes #12.Try it
🤖 Generated with RuFlo
Co-Authored-By: RuFlo ruv@ruv.net