Plugin Claude Code
Plugin marketplace listing 1 plugin: vedix.
3 3mo ago A
tokens not measured
original MIT
vedix
02Plugin Claude Code
Vedix — research workbench that turns a topic into a venue-ready manuscript via cross-host CLI orchestration.
3 3mo ago A
tokens not measured
original MIT
Agent
Iteratively adds missing citations to references.bib (up to 5 rounds). Searches Semantic Scholar by topic gap. Never fabricates metadata.
3 3mo ago A 34 tokens
original MIT
Agent
Generates a runnable Python experiment script (experiment.py) plus its requirements.txt, implementing the methodology from hypothesis.md using only the domain template's preferred libraries.
3 3mo ago A 36 tokens
original MIT
Agent
Scans a target codebase to extract entry points, modules, dependencies, key patterns, and extension points. Emits structured codebaseanalysis.json. Triggered when user provides --codebase or asks to analyze/audit a repo.
3 3mo ago A 53 tokens
original MIT
Agent
CLAUDE-CODE-EXCLUSIVE meta-agent. Cross-validates a Claude phase output against Codex via the codexbridge CLI. Returns a structured verdict (agree/minordisagree/majordisagree) plus discrepancies and Codex's alternative. The orchestrator dispatches this after every cross-validatable phase (ideation, hypothesis…
3 3mo ago A 101 tokens
original MIT
Agent
Installs requirements.txt into the per-job venv, runs experiment.py with timeout, parses stderr on failure, patches the script up to 3 rounds. Returns a structured run report.
3 3mo ago A 44 tokens
original MIT
vedix-fixer
08Agent
Diagnoses pipeline failures (network, dependency, schema-mismatch, runtime, timeout, output-parse) and returns 2-4 concrete fix options for the orchestrator to surface to the user via AskUserQuestion. Never auto-applies fixes silently.
3 3mo ago A 56 tokens
original MIT
Agent
Produces a testable hypothesis with mathematical models, statistical framework, methodology, and codebase integration plan. Grounded in literature + prior meta-analysis.
3 3mo ago A 36 tokens
original MIT
Agent
Generates a structured research idea (Name, Title, Hypothesis, Related Work, Abstract, Experiments, Risks) with novelty check via OpenAlex/Semantic Scholar. Reads prior meta-analysis to avoid re-treading failed approaches.
3 3mo ago A 53 tokens
original MIT
Agent
Per-source literature search worker. The orchestrator dispatches this agent ONCE PER SOURCE in parallel (up to 6 sources), each invocation querying only its assigned source via the source's dedicated MCP server. Returns a normalized paper list for that one source. Orchestrator merges across all returns.
3 3mo ago A 66 tokens
original MIT
Agent
Writes a complete LaTeX scientific manuscript by orchestrating 6 nested section subagents in parallel (Abstract, Introduction, Methods, Results, Discussion, Conclusion). Enforces consistency (citations, figure refs, no contradictions, no placeholders). Picks LaTeX template per settings.
3 3mo ago A 64 tokens
original MIT
Agent
Reads all trajectories and jobs, computes per-domain success rates and failure patterns, writes metaanalysis.json + whatworks.json with concrete recommendations for future jobs.
3 3mo ago A 38 tokens
original MIT
Agent
Agent "paper-extractor" from danilkotelnikov/vedix, covering required output structure, hard constraints and output format.
3 3mo ago A 0 tokens
original MIT
Agent
Generates publication-grade figures via 3-cycle iterative refinement. Two output paths: matplotlib/seaborn scripting and LaTeX-native TikZ/pgfplots. Per-cycle artifacts in figuresdraft1/, figuresdraft2/, figuresfinal/. Auto-applies Okabe-Ito palette and journal-specific rcParams. Uses VLM critique on a 10-point rubric…
3 3mo ago A 85 tokens
original MIT
Agent
Performs NeurIPS-format peer review of the manuscript. Multimodal — also performs visual validation pass on rendered PDF/DOCX pages. Produces review.json + manuscriptv2.tex with top-3 fixes applied.
3 3mo ago A 50 tokens
original MIT
Agent
Generates Beamer PDF + python-pptx editable slide deck + speaker notes from a compiled manuscript. Vendored from AI-Research-SKILLs presenting-conference-talks. Runs after Phase 7 (reviewer accepts).
3 3mo ago A 53 tokens
original MIT
Agent
Best-First Tree Search (BFTS) experiment runner. Wraps the canonical Sakana AI-Scientist treesearch/performexperimentsbftswithagentmanager.py module. Explores multiple experiment-implementation variants in a tree structure, parallel-evaluates them, and picks the best by metric. 5-20x slower than single-shot…
3 3mo ago A 103 tokens
original MIT
Agent
Vision-language review of rendered manuscript figures. Wraps the canonical Sakana AI-Scientist performvlmreview.py module. Detects duplicate figures, validates captions match content, scores each figure for clarity/relevance/quality, and flags figures that don't match their referenced text in the manuscript.…
3 3mo ago A 100 tokens
original MIT
vedix-list
20Command
List all Vedix research jobs.
3 3mo ago A 7 tokens
original MIT
vedix-meta
21Command
View the Vedix cross-job meta-analysis.
3 3mo ago A 9 tokens
original MIT
vedix-output
22Command
Retrieve a section of a completed Vedix job's output.
3 3mo ago A 12 tokens
original MIT
vedix-query
23Command
Search the Vedix persistent knowledge store (papers, hypotheses, benchmarks, claims).
3 3mo ago A 16 tokens
original MIT
vedix-resume
24Command
Resume a paused or failed Vedix job from its last successful phase.
3 3mo ago A 14 tokens
original MIT