obzenner

obzenner/siffran

Plugin Claude Code

Dmitry's personal Claude Code plugin marketplace — methodology-driven tools for rigorous engineering.

obzenner/siffran

Skill Claude CodeCodex

Run a structural health check on this marketplace repo and regenerate the plugin tables in CLAUDE.md and README.md. Verifies manifest integrity, cross-references, version hygiene, then rewrites the generated documentation regions from source manifests. Use before committing, after adding plugins/methodologies or…

obzenner/siffran

Instructions file

Claude Code instructions for obzenner/siffran, covering claude.md, drive this project through the makefile, repo structure, conventions and versioning.

obzenner/siffran

Plugin Claude Code

Empirical-convergence workflow — adjudicates a claim graph (GSN argument with in-toto evidence) where every claim's confidence must be earned by real external evidence: research citations first, deterministic spike verdicts for machine-checkable claims, then an independent auditor on a different model before a run may.

obzenner/siffran

Agent

Independent auditor for an empirica run. Verifies the run against the ADR-20 rubric — re-reading each approved claim's research citation to confirm the cited source actually supports the claim — and writes a pass/fail verdict artifact. Spawn this before a run may report converged. The authoring agent must never write…

obzenner/siffran

Agent

Fold-1 research worker for an empirica run. Resolves a needs-data claim by fetching and citing a real source outside the model's training data, then returns a structured research record. Use for the mechanical fetch-and-cite work of the empirical loop.

obzenner/siffran

Agent

Fold-2 spike worker for an empirica run. Designs a real deterministic check for a needs-experiment claim, runs it through spikeharness.py so the verdict is a genuine exit code, and reports the result. Use when a claim needs runtime evidence rather than a citation.

obzenner/siffran

Hook

Runs to expand your prompt before the agent sees it for ^empirica:empirica$ tool calls, running python3. From obzenner/siffran.

obzenner/siffran

Hook

Runs before the agent uses a tool for Agent, Read, Glob, Grep, Bash, WebFetch, WebSearch, NotebookRead, LSP and Bash tool calls, running python3 (3 commands). From obzenner/siffran.

obzenner/siffran

Hook

Runs when the agent finishes a response, running python3. From obzenner/siffran.

obzenner/siffran

Hook

Runs when a session starts on compact, running python3. From obzenner/siffran.

obzenner/siffran

Skill Claude CodeCodex

Empirical-convergence development workflow. Adjudicate a claim graph — propose claims, earn each one's confidence with real external evidence (research first, then a deterministic spike where the claim is machine-checkable), discard what evidence refutes — then have an independent auditor verify the run before it may…

obzenner/siffran

Plugin Claude Code

Methodology router — picks and executes formal CS/math reasoning methodologies with tracked phases and structured output.

obzenner/siffran

Skill Claude CodeCodex

Select and execute a formal reasoning methodology for the current task. Use when facing architectural decisions, debugging, rule enforcement, design tradeoffs, assumption validation, or any situation requiring structured thinking. Trigger phrases: 'think through this', 'reason about', 'which approach', 'analyze this…