Defense opportunity intelligence · MCP
Watch the federal pipeline. Score it against your company. Act from inside your agent.
DefScout is a remote MCP server. It continuously ingests SAM.gov, DoD SBIR/STTR (DSIP), and Grants.gov, ranks every solicitation against your capability profile with Claude, and keeps pursuit state (threads, overrides, digests) conversational.
Why an MCP instead of another dashboard
Capture managers already live in chat and code. DefScout meets them there: ask for today's digest, pull attachment text, record a founder YES/NO with a structural reason, estimate ROM cost without context-switching into a SaaS portal that goes stale.
- 01 · Corpus stays warm (SAM CSV + API, DSIP, Grants) with hybrid search and local vectors.
- 02 · Scoring is company-specific: products, NAICS, watch offices, policy overrides.
- 03 · Pursuit is stateful: threads, engagement events, deadlines, email digests.
Fifty-six tools. Four jobs.
Grouped the way a capture team works, not the way an API catalog lists endpoints.
Discover
- search_sam / lookup_solicitation
- search_dsip / search_grants
- get_daily_digest / get_daily_arxiv
Triage
- get_solicitation_detail + attachments
- override_score with structural reasons
- get_position / estimate_cost
Pursue
- pursue list + engagement timeline
- thread watches & lineage
- DEFCON alerts for deadline pressure
Respond
- company_info boilerplate
- response_templates by vehicle
- capability profile + watch terms
In the terminal of your agent
you What should I look at before standup?
defscout Calling get_daily_digest…
Top 5 above threshold: AFWERX Open Topic (82), ONR Autonomy BAA amendment (79), Army C5ISR recompete draft (74)…
you Pull the AFWERX attachments and tell me if our OMS work is in scope.
defscout list_solicitation_attachments → get_attachment_content · windowed match on “open architecture”…
Ready to wire it up? Sign in to authenticate