Tools provided
Tools the agent exposes — your AI client calls them automatically when it needs them.
- reconcile_records — Line two sets of records up and report what matched YOUR rule, what only nearly did, and what matched nothing on either side — with the difference always stated, because that is the number a controller signs off. A near miss is reported with the exact amount difference and the exact number of days, never a confidence: the tool this replaces scored pairs 0.95 / 0.85 / 0.65 / 0.6 and auto-reconciled anything over 0.9.
- read_aging — Age every open item against a date YOU supply: days overdue on each (negative when not yet due), totals per bucket from YOUR edges, and where the balance is concentrated by counterparty. No risk score, no suggested action and no collection probability — the tools this replaces mapped each bucket to a probability of being collected and blended a score from weights of 0.5, 0.3 and 0.2 to decide which customer gets chased.
- read_variance — Compare actual against budget line by line: the variance, the percentage, and which way it went — largest first. Favourable or adverse is reported ONLY for a line that says whether it is revenue or cost, because over budget is good news on one and bad news on the other; the tools this replaces called every over-budget line favourable, which is backwards for every expense line. Nothing is called material unless YOU supply the rule.
- 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.
