Tools provided
Tools the agent exposes — your AI client calls them automatically when it needs them.
- read_clause_coverage — Say which of YOUR standard clauses appear in a contract and where, quoting the span around each one so you read the clause rather than a label. A clause that matched nothing is reported as NOT FOUND together with the phrases looked for and how much text was searched — never as absent, and never as safe. No risk score: the tool this replaces multiplied category weights by a 0.4/0.7/1 severity multiplier into a number out of 100.
- check_playbook_deviation — Say where a contract departs from YOUR playbook, quoting the clause, your preferred position and the phrase that decided it. THREE verdicts, not two: a clause can depart from your position, satisfy it, or address it NEITHER WAY — and the third is a question, not a redline. A clause no rule governs is unresolved, and a rule no clause matched is named as a gap in the review. Severity is yours, reported exactly as your rule declared it.
- read_obligation_calendar — Quote every sentence in a contract that creates a duty, report the date or period it states, and say how many days are left — measured from a date YOU supply, and NEGATIVE where the date has already passed. Ordered soonest first, so an overdue deadline leads. No urgency band: the tool this replaces classified days into bands, mapped those to a severity and derived a recommended action, so every layer after the band inherited the invention.
- check_control_evidence — Say which controls of YOUR framework have nothing filed against them, naming every gap by id — and which evidence names no control at all, which every tool this replaces dropped silently. Reported as COUNTS, never a coverage percentage or a maturity score out of 100: "82% ready" is a sentence no auditor agrees with, and both numbers were invented.
- run_full — Run every capability you supplied the input for, in registry order, and return one report. Capabilities whose input is missing are listed as skipped with the fields they need — never run against a fixture and never reported as zero.
- list_capabilities — List what this agent can do: every tool it advertises, what each one is for, and the arguments each one reads. Useful for discovery before the first call.
- plan_inputs — Plan the inputs for a tool: returns the questions to ask, the JSON schema of the arguments, and a ready-to-edit example. Every argument it names is an argument the tool actually reads.
- discover_intent — Understand your goal and co-design the exact input through clarifying questions before anything runs.
- open_form — Use this the moment the user wants to run this agent on their own input. It opens an interactive form that collects the run arguments. Do NOT check credentials, do NOT ask clarifying questions in chat, and do NOT run any analysis before calling this — the form is the entry point for running the agent.
- run_form — Internal: invoked by the open_form form when the user submits. Reads exactly the same arguments as run_full and forwards all of them. Not for direct use.
