theorems-harness-product

Plugin theorems-harness-product from marketplace theorems-harness-marketplace in Travis-Gilbert/Theorems-Harness. Commands read from the repo's manifest — 27 days ago.

Skills (16)

• compute-offload

  Use when the user asks a graph or tabular question with an exact answer (reachability, transitive closure, shortest or weighted paths, counts, aggregations, set operations, joins), or mentions connectors, external tools, invoke, affordances, or "which tool". Route exact symbolic work to compute_offl

• compute-graph-plans

  Chart a destination as a traversable plan graph on the Theorem substrate, ending at a gate. Harness binding: requires the Theorem MCP surface (plan tool, multihead_*). Agents without those tools — or with a listed-but-dead plan surface — use the portable binding of this skill.

• dispatch

  The local-loop job board and session spawn surface. Use when the user wants to hand a task from a planning surface to an executing head, queue work for later, read or update the job board, close a job, wake an asleep head, or start a fresh Claude Code or Codex session locally. Triggers on "queue a j

• execute-graph-plans

  Traverse a compiled plan graph on the Theorem substrate, one move per turn, until fixpoint. Harness binding: requires the Theorem MCP surface (plan tool with the motion actions, multihead_*, coordination_*, encode, continuity_pack). Agents without those tools use the portable binding of this skill.

• execute

  Carry a bounded task through code changes and validation inside a receipted run. Use when the user asks to implement, fix, ship, simplify, run tests, reconcile a plan or checklist, merge, or deploy; when a run surprises you mid-flight; or when a claim of done needs evidence. Claims from /planning-th

• harness-coordinate

  Teach and run Theorem's Harness cross-agent coordination protocol. Use when Codex, Claude Code, Claude.ai, or another head works the same repo/task and the user wants coordination, shared intent, mentions, presence, handoff, or a ping-like wait. The model is one agent with several heads sharing a sc

• index

  This skill should be used when the user asks to "index context", "query memory", "find prior decisions", "assemble context", "inspect adaptive index state", "explain retrieval receipts", or needs memory, code maps, query receipts, and index-aware context before answering or acting.

• planning-theorem

  Chart a fuzzy ask into a durable Plan on the Theorem substrate, then work its decision frontier one decision per session. Use when an idea arrives wrapped in fog, when work spans more than one session or head, when acceptance criteria are not yet locked, when the user asks for a plan, map, spec, mig

• prototype-theorem

  Raise the fidelity of a decision by building a throwaway artifact to react to; a logic spike, a set of UI variations, an outline, or a stub. Use when a Plan decision task hinges on how something should look or behave, or when discussion has gone abstract and needs something concrete to push against.

• replay-last-run

  Show the event timeline of a Theorem's Harness run or the bounded replay of a plan. Use when the user asks "what did you actually do," "show me the run trace," "walk me through the run," "audit what happened," "what tools fired and in what order," "why was this task refused," or "what happened on th

• research-theorem

  Resolve an AFK research task from a Plan against primary sources, offload the findings to the graph, and land a markdown projection in the repo. Use when a decision waits on facts from documentation, third-party APIs, source code, benchmarks, or the knowledge graph; dispatched by /planning-theorem f

• reverse-engineer

  This skill should be used when the user asks to "reverse engineer" a repository, URL, website, codebase, product surface, feature, API, workflow, or binary artifact, or asks to "reconstruct" behavior into a grounded map, behavior spec, parity checks, or rebuild plan.

• rust-engineering

  This skill should be used when the user asks to "write Rust", "debug Rust", "review Rust", "fix Cargo", "work on a Rust MCP server", "improve a Rust crate", or needs help with Cargo workspaces, Rust validators, PyO3/FFI bridges, async Rust services, parsers/macros, systems code, Ensemble pack select

• spec-theorem

  Compile a specification as a method node inside a Plan; the artifact plus its obligation set, with every spec section covered by at least one plan task. Use when a Plan's destination is a spec, when a resolved decision calls for a specification deliverable, or when the user asks for a SPEC or HANDOF

• theorems-harness

  Theorem's Harness default plugin skill. Use when the user invokes /harness or asks for grounded planning, implementation, debugging, review, context preparation, validation, reporting, cross-agent coordination, typed memory, encode, run lifecycle, replay, fork/compare, fractal search, code search, o

• theorize

  Turn fuzzy intent into a named destination and a contested option space before the wrong implementation begins. Use when the task is hedged or fuzzy ("could we", "maybe", "what if"), when multiple credible approaches exist and the wrong choice causes real churn, when a design decision is contested,

Slash commands (5)

• /theorems-harness-product:code

  Query or ingest code context through compute_code, understand_code, impact, and oracle.

• /theorems-harness-product:graphql

  Run Theorems GraphQL reads, mutations, or schema introspection through the product MCP facade.

• /theorems-harness-product:grep

  Search files, code neighborhoods, memory, or web evidence through grep, semantic_grep, memory_grep, and mgrep.

• /theorems-harness-product:memory

  Retrieve memory and evidence through query_data, retrieve_memory, turn_start, and evidence_bundle.

• /theorems-harness-product:reconstruct

  Reconstruct a repo, URL, API, feature, workflow, or binary into a grounded spec or component plan.

MCP servers (2)

• theorems-harness-product

  runs: sh -lc cd "${CLAUDE_PLUGIN_ROOT:-${PLUGIN_ROOT:-.}}" && exec node src/mcp/server.mjs

• theorems-harness-remote

  connects to: https://travis-theorem-personal.fly.dev/mcp

Hooks (9)

• PostToolUse

  node "${CLAUDE_PLUGIN_ROOT}/src/bin/theorem-hook.mjs" claude_code

• PostToolUseFailure

  node "${CLAUDE_PLUGIN_ROOT}/src/bin/theorem-hook.mjs" claude_code

• PreToolUse

  node "${CLAUDE_PLUGIN_ROOT}/src/bin/theorem-hook.mjs" claude_code

• SessionEnd

  node "${CLAUDE_PLUGIN_ROOT}/src/bin/theorem-hook.mjs" claude_code

• SessionStart

  node "${CLAUDE_PLUGIN_ROOT}/src/bin/theorem-hook.mjs" claude_code

• Stop

  node "${CLAUDE_PLUGIN_ROOT}/src/bin/theorem-hook.mjs" claude_code

• Stop

  bash "${CLAUDE_PLUGIN_ROOT}/hooks/wiring-review-stop.sh"

• StopFailure

  node "${CLAUDE_PLUGIN_ROOT}/src/bin/theorem-hook.mjs" claude_code

• UserPromptSubmit

  node "${CLAUDE_PLUGIN_ROOT}/src/bin/theorem-hook.mjs" claude_code

Similar plugins