Kriday Dave
Community@kridaydave
Im Kriday. Founder of @Epoch-AI-Lab
Agent Skills by Kriday Dave
Showing 31 vetted skills indexed across 2 GitHub repositories.
kridaydaveswarm
Spawns parallel blind subagents to exhaustively search a corpus and unions their deduplicated findings.
kridaydavevera
Synthesizes research evidence and analyzes datasets into hypotheses and recommendations.
kridaydavejury
Verifies a single claim by spawning parallel subagents that attempt to refute it and tallying their votes.
kridaydaveecho
Searches crew memory and inbox files with parallel subagents and returns deduplicated matches.
kridaydaveorator
Designs narrative arcs, slide-by-slide plans, and Q&A prep for talks and presentations.
kridaydavecouncil
Dispatches parallel blind subagents embodying personas and diffs their structured verdicts into a disagreement map.
kridaydaveprism
Designs publication-quality charts and figures by matching data shape to honest visual encodings.
kridaydavecodex
Extracts contributions, methods, evidence, assumptions, and limitations from a single research paper into structured notes.
kridaydavetournament
Generates divergent solution candidates in parallel and judges them through a bracket to synthesize a winner.
kridaydavecipher
Writes, debugs, and evaluates LLM prompts using structured prompt-engineering techniques.
kridaydavecrew-status
Reads crew memory and inbox files to report persona knowledge and pending handoffs.
kridaydavetrove
Builds datasets through sourcing, sampling, labeling, deduplication, splitting, and datasheet documentation.
kridaydavefind-skills
Searches and installs agent skills from the open skills ecosystem using the Skills CLI.
kridaydavecrew-init
Scaffolds the crew memory directory structure with agents and inbox folders.
kridaydavedebate
Orchestrates structured adversarial debates between two AI personas to stress-test plans and decisions.
kridaydaveethos
Reviews research studies for ethical risk, consent scope, bias, and regulatory compliance.
kridaydaveforge
Writes, debugs, and reviews code with systematic root-cause analysis and verification.
kridaydavecrew-sync
Audit crew memory files for orphans, stale entries, duplicates, and format errors.
kridaydavecharter
Writes PRDs, user stories, and prioritized backlogs from vague product ideas.
kridaydavegauge
Designs evaluation metrics, benchmarks, and significance tests to validate model improvements.
kridaydavescribe
Writes and revises research papers, technical documentation, and proposals with evidence-matched claims.
kridaydavecompass
Plans project timelines, dependencies, and critical paths for multi-week work.
kridaydaveanchor
Pin environments, seeds, and data versions to build reproducible replication packages.
kridaydavelemma
Proves, disproves, and verifies mathematical claims with step-by-step justified derivations.
Frequently Asked Questions About Kriday Dave
FAQPage SchemaWhat tasks can I accomplish with Kriday Dave's skill registry?▼
The 31 skills cover the full research lifecycle: literature scouting (Beacon), deep paper reading (Codex), experiment design (Helix), dataset construction (Trove), data analysis (Vera), mathematical proofs (Lemma), adversarial verification (Jury, Quill, Debate), scientific writing (Scribe), figures (Prism), talks (Orator), and reproducibility packaging (Anchor).
Who is the target audience for these skills?▼
Researchers, data scientists, ML engineers, and technical founders. Personas like Forge (senior engineer/debugger), Atlas (system architect), Charter (product manager), Compass (program manager), and Gauge (evaluation/benchmarking) also serve software teams needing design, prioritization, and measurement rigor.
How do the multi-agent orchestration skills work in practice?▼
Aleth auto-routes any query to the right persona chain. Council polls blind parallel personas and diffs disagreement; Swarm fans out finders for exhaustive search; Tournament brackets competing solutions; Jury refute-votes claims. Crew-init scaffolds memory/agents/ and memory/inbox/ directories for persistent cross-persona handoffs.
What are the prerequisites for using the crew memory features?▼
Run crew-init once per project to scaffold memory/agents/ and memory/inbox/ directories plus a README convention. Echo, crew-status, and crew-sync then search, dashboard, and audit that memory layer read-only. The existing file-organizer-dev skill additionally requires Zod schema and path-validation patterns for MCP server development.
Do these skills support research ethics and compliance review?▼
Yes. Ethos handles IRB preparation, informed consent design, bias and fairness audits, PII/GDPR/HIPAA privacy checks, conflict-of-interest surfacing, and dual-use risk flagging. Trove separately covers dataset licensing, provenance, and datasheets to keep data collection compliant before analysis begins.