caveman
claudeTalk like caveman. Cut filler. Keep technical details explicit.
22 skills · 1 command · 3 agents · hooks
Generate synthetic M&A virtual data rooms with a verified ground-truth answer key, for evaluating document-review and due-diligence AI.
Plugin synth-vdr from marketplace synth-vdr in gblegal/gb_synth_vdr. Commands read from the repo's manifest — 13 days ago.
vdr-probeRun the vdr-build probe, which checks whether this Legora workspace has what a synthetic data-room build needs — sub-agents that can write files and start with a clean context, a scratch folder that lasts, zips that go out and come back intact, and shell calls long enough to finish. Use when asked t
explain-a-planted-findingAnswer questions about what was planted in a synthetic test data room, such as "what was planted in this document", "where is CORP-1", "was that a red herring" or "what does ENV-2 rest on", by quoting the room's answer key. Use when adjudicating a review of a synth-vdr room. Reads the key only, and
review-a-test-data-roomReview a synthetic test data room as a buyer's lawyer would, and hand back every issue found, each with its severity and the documents it rests on, as a saved report. Use when asked to review a test data room, run due diligence on a synth-vdr room, or list the issues in a benchmark room. For synthet
vdr-buildAuthor the data room in waves of subagents, consolidating answer-key refinements and running the QA gates after every wave. Resumable — re-running continues from the build status file rather than starting the room over.
vdr-findingsDesign the findings registry, distractors and evidence-chain map for a scoped data room — the eval specification, fixed before a single document is authored. This is a hard gate; no authoring proceeds until the registry is signed off.
vdr-packagePackage a finished data room — build the deterministic subset, render optional DOCX/PDF trees, write the content manifest and freeze the room for release. Runs every QA gate in --strict mode; refuses to freeze a room with any failure or any skip.
vdr-qaRun the twenty-two room QA gates — index regeneration, leakage sweeps, twin invariants, carrier census, cross-references, depth lint, subset, fact-sheet and answer-key reconciliation, unchecked-name sweep, discoverability, render parity, the room's exemplar/eval role declaration, the eval room's cla
vdr-scopeScope a new synthetic M&A data room — a very light-touch interview, an invented deal fact sheet, room.conf, the slot index, and a name-collision check that blocks Gate A on any unresolved collision. Run this first, before any other /vdr- command.
vdr-scoreScore a document-review tool's output against the room's answer key — recall by severity, precision, false alarms against the distractors, partial credit on multi-document trails, provenance verification, and a baseline diff between two tool runs. Also scores a classification tool's output against t
vdr-auditorReads the blind room only and determines whether each planted finding is genuinely discoverable from it, returning its verdict for /vdr-build to record in the answer key. Used by /vdr-build after authoring completes. Read-only — never writes anywhere itself.
vdr-authorAuthors a batch of data-room documents blind-first — the natural seller-side document plus its answer-key refinement — and returns a wave manifest. Used by /vdr-build for wave fan-out. Never writes the flagged tree.
Talk like caveman. Cut filler. Keep technical details explicit.
22 skills · 1 command · 3 agents · hooks
Persistent memory for Claude Code: the next session starts where the last one left off.
29 skills · 1 MCP server · hooks
tasteskill 2.5.1: professional company websites at the figma.com and x.ai level, with one section per screen, real photographs, real product UI, effects kept for tech and gaming pages, and a real-browser checker and finish gate.
1 skill
Open-source AI job search that runs locally in your AI coding CLI: scan job portals, evaluate listings into a structured A-H report with a 1-5 score, tailor your CV, track applications
5 skills · hooks
Cross-session memory and token savings for coding agents.
6 skills · 1 agent · 1 MCP server · hooks
Package any website or local web build into a lightweight desktop app with Pake CLI.
1 skill
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.