Multi-Agency Forecast Surveillance Agent

Autonomously watch 30+ agency forecasts + SAM.gov new opportunities + USAspending obligation deltas on a weekly cadence (default Monday 06:00 ET, user-overridable) and produce a synthesized intel digest tailored to a pasted capability profile + target-agency list + saved searches. The agent runs 7 sub-agents in sequence — Parser, Forecast Watch (wraps Tool #25), New Opportunities (lateral Tool #3), Obligation Delta (net-new USAspending integration), Synthesis (wraps Tool #26 for capability-match scoring), Recommendations (net-new feedback loop), Digest Assembly — and produces a weekly Markdown digest with top-N opportunities, 4-state forecast-to-actual signals (ON TRACK / SLIPPING / MATERIALIZED / GHOSTED), obligation deltas, recommended saved-search refinements, and an idempotent state diff for next-run consumption. Collapses 8-16 hrs/month of manual forecast monitoring into a 15-30-minute autonomous weekly run that compounds in value via the capability-profile + saved-search refinement loop.

What you get

A production-ready Claude agent that runs the workflow above on your own subscription — no SaaS contract, no procurement cycle, no third-party data layer. Install it and run it on the next pursuit that hits your desk.

How it ships

Delivered as installable source from Products/agents/multi-agency-forecast-surveillance, with an install guide, a verification step, and a published scope (what it does, what it doesn't). Runs inside your Claude instance on the data you provide.

HIGH-TRUST discipline. Where this tool produces pricing, compliance, or past-performance artifacts, FAR/DFARS guardrails and "FOR INTERNAL ESTIMATE USE — NOT A PROPOSAL PRICE" disclaimers are embedded in the output. It does not replace qualified pricer, counsel, or color-team review.

1.0.0

• Initial catalog release.