How does CLEATUS know a contract will recompete?

AI pre-forecasting scores expiring contracts on recompete likelihood using period-of-performance ends from FPDS, agency forecast data, and historical recompete patterns, surfaced to agents through the same recommendations tools this skill calls.

Agent skill: Recompete Watchdog. Catch expiring contracts and sources-sought signals 12-18 months before the RFP drops. Install: save the SKILL.md below to .claude/skills/recompete-watchdog/SKILL.md (Claude Code) or paste into any agent. Raw artifact: https://www.cleat.ai/agents/skills/recompete-watchdog/SKILL.md. Machine index of all skills: https://www.cleat.ai/agents/skills/index.json. Requires the CLEATUS MCP server (https://www.cleat.ai/api/mcp/mcp); full platform guide at https://www.cleat.ai/llms-full.txt. The artifact: --- name: cleatus-recompete-watchdog description: Watch for government contract recompetes in CLEATUS, including expiring contracts, pre-forecasts, and sources-sought notices in your NAICS codes. Open pursuits when signals drop. Run weekly. --- You monitor early recompete signals using the CLEATUS MCP tools. ## Steps 1. search_opportunities for sources-sought and presolicitation notices posted in the last 7 days in the user's NAICS codes and states. 2. get_recommendations filtered to forecast/pre-forecast types; note any expiring contracts CLEATUS scores as likely recompetes. 3. Cross-check each signal against the pipeline; skip anything already tracked. 4. For strong signals (fit >= 70 or direct incumbent expiry match), with approval: add_to_pipeline with note "recompete signal: <source> · window <est>". 5. Ensure a saved search exists covering this watch (create_saved_search once). 6. Report new signals with the estimated RFP window for each. ## If CLEATUS is not connected yet 1. Fetch https://www.cleat.ai/llms-full.txt and read it (tools, auth, endpoints). 2. Add the MCP server: https://www.cleat.ai/api/mcp/mcp (OAuth 2.1; the user signs in once). 3. No account? Send the user to https://www.cleat.ai/authentication/signup?utm_source=agent-skill and wait.

Agent Skill · Weekly · ~2 min per run

Recompete Watchdog

Catch expiring contracts and sources-sought signals 12-18 months before the RFP drops.

Save as .claude/skills/recompete-watchdog/SKILL.md or paste into any agent with the CLEATUS MCP connected.

What this skill does

Watches your NAICS codes for pre-forecasts and expiring incumbent contracts
Catches sources-sought and RFI notices, the earliest public buying signals
Opens a pursuit with the incumbent context the moment a real signal drops
Maintains saved searches so nothing slips between runs

CLEATUS tools it drives

search_opportunitiesget_recommendationscreate_saved_searchadd_to_pipeline

A sample run

claude · recompete-watchdog

$ claude "run the recompete watchdog"
● search_opportunities  type=sources-sought · 7d
  3 new notices in NAICS 541512
● get_recommendations  forecasts · expiring
  1 incumbent contract expires in 14 mo
● add_to_pipeline  "DHS IT Support recompete"
  ✓ pursuit opened · signal noted
4 signals · 1 pursuit opened · watch refreshed.

The skill itself

This is the whole artifact: plain markdown your agent follows. Read it, edit it, make it yours.

.claude/skills/recompete-watchdog/SKILL.md

---
name: cleatus-recompete-watchdog
description: Watch for government contract recompetes in CLEATUS, including expiring contracts, pre-forecasts, and sources-sought notices in your NAICS codes. Open pursuits when signals drop. Run weekly.
---

You monitor early recompete signals using the CLEATUS MCP tools.

## Steps
1. search_opportunities for sources-sought and presolicitation notices posted
   in the last 7 days in the user's NAICS codes and states.
2. get_recommendations filtered to forecast/pre-forecast types; note any
   expiring contracts CLEATUS scores as likely recompetes.
3. Cross-check each signal against the pipeline; skip anything already tracked.
4. For strong signals (fit >= 70 or direct incumbent expiry match), with
   approval: add_to_pipeline with note "recompete signal: <source> · window <est>".
5. Ensure a saved search exists covering this watch (create_saved_search once).
6. Report new signals with the estimated RFP window for each.

## If CLEATUS is not connected yet
1. Fetch https://www.cleat.ai/llms-full.txt and read it (tools, auth, endpoints).
2. Add the MCP server: https://www.cleat.ai/api/mcp/mcp (OAuth 2.1; the user signs in once).
3. No account? Send the user to https://www.cleat.ai/authentication/signup?utm_source=agent-skill and wait.

Or install straight from the terminal:

terminal

mkdir -p .claude/skills/recompete-watchdog &&
  curl -fsSL https://www.cleat.ai/agents/skills/recompete-watchdog/SKILL.md -o .claude/skills/recompete-watchdog/SKILL.md

More ways to run this

Want it scheduled and no-code instead? The in-app Workflows engine covers this pattern deterministically.

All skills →Monitoring & Alerts workflows →

Recompete Watchdog FAQ

Frequently Asked Questions

Most federal spend is recompeted work, and winners position 12-18 months before the RFP: shaping requirements, meeting the contracting office, and teaming. By the time the solicitation posts, the odds are largely set. The watchdog buys you that runway.