Audit one bounded Jacobian mathematical slice for missing or unusable executable vocabulary using source demand, catalog composition, contract evidence, and maintained-backend feasibility. Use for proactive gap audits that are broader than one operation but narrower than a repository-wide inventory; do not use for…
Audit existing MCP tools for execution friction after an agent has reached the tool. Use when reviewing public JSON Schemas, Pydantic validation, operation examples, error recovery, completed agent traces, or benchmark runs to explain malformed calls, retries, or unusable results. Do not use to measure whether agents…
Audit new or existing Jacobian mathematical operations for public-domain mismatches, hidden work expansion, evidence-backed scale or backend improvements, lossy exact results, source-unbound conclusions, and producer-consumer incompatibility. Use for operation-contract reviews, mathematical performance investigations…
Decompose a bounded corpus of mathematical proofs, formalizations, scripts, and certificates into recurring solution techniques and the smallest reusable Jacobian postconditions. Use for repository- or corpus-level “what can Jacobian learn?” audits; do not use for one operation contract or one agent trajectory.
Diagnose whether agents can see, discover, select, and correctly execute MCP tools without conflating those stages. Use when investigating why agents did or did not use an MCP tool, evaluating MCP server instructions or tool descriptions, testing natural-language discovery, or designing a frozen control/treatment…
Build, validate, and run Jacobian evaluations packaged as Harbor datasets. Use when authoring or changing Harbor tasks, independent verifiers, Oracle jobs, workflow fixtures, task digests, or evaluation handoffs.
Use when solving nontrivial mathematical problems or investigating conjectures that may benefit from Jacobian's local library of typed operations for exact computation, finite search, structural analysis, formal reasoning, probability, or optimization.
Review completed or paused mathematical agent transcripts, visible reasoning, code, searches, tool calls, corrections, and final claims to extract evidence-backed lessons for Jacobian operations, discovery, contracts, skills, evaluations, and documentation. Use for mathematical workflow retrospectives and "what should…
Design, audit, and repair fail-closed mathematical verifiers and evaluation contracts, including public schemas, frozen-input binding, task-specific witnesses, diagnostic scoring, adversarial fixtures, and Oracle validation. Use when a verifier can crash, accept malformed claims, reject equivalent witnesses, leak…