# SentinelPulse

- **Event:** [Meta Global AI Developer Hackathon](https://cerebralvalley.ai/e/meta-ai-developer-hackathon)
- **When:** Oct 26, 12:00 PM – Nov 4, 9:00 PM (EST)
- **Where:** Online
- **Team:** [AMIT KUMAR MISHRA](https://cerebralvalley.ai/u/sam8873)
- **GitHub:** https://github.com/bb8654838-lgtm/sentinelpulse
- **Demo video:** https://youtu.be/rkK1-N0POcQ?si=_c6OpxBA57AjV4GI
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/46

Save model reasoning for alerts that need it. A tier-0 gate applies deterministic rules to suppress known noise at zero tokens and escalates unresolved alerts to Meta Muse (muse-spark-1.3), which turns each escalation into a triage note with hypotheses and next actions. On a 7-fixture replay, gate off vs on: 0 noise model calls, -13.2% tokens, -19.8% latency, 0/7 incidents missed.

Every on-call queue fills with non-incidents: health checks returning 200, favicon 404s, CDN revalidations, blips that self-recovered before anyone looked. Most "AI SRE" stacks send it all to a reasoning model, paying in tokens, latency and attention for nothing.

SentinelPulse puts a deterministic gate before the model:
- gate-0 applies rules to the normalized event, dropping provable noise at zero tokens.
- A cheap classifier categorizes the survivors.
- Only events that clear gate-0 reach tier-1 triage on Meta's Muse API (muse-spark-1.3), which drafts the triage note, hypotheses and next actions.
- A sandboxed remediator can propose a minimal patch; the authenticated PR path is opt-in and unused in the demo.

The gate does not make the model smarter; it keeps noise away from it. Same 7 fixtures replayed, gate off vs gate on, single run:
  C1 incidents missed      0/7 -> 0/7
  C2 noise model calls     4   -> 0     (all four blocked by rules)
  C3 tokens                5533 -> 4802 (-13.2%)
  C4 end-to-end latency    36.8s -> 29.5s (-19.8%)
  C5 signal escalation     3/3 -> 3/3

SHA-256 freezes the evidence (bench/evidence/FREEZE-20261027.json), tying every number to exact bytes. `python3 -m demos.l3.freeze_evidence --check` fails loudly on any drift, and `bash demos/l3/demo_take.sh` runs a 4-step preflight gate. The report is a self-contained, CDN-free HTML file rendering air-gapped. Live mode refuses with exit 4 without --confirm-live, and BudgetGuard reserves calls and tokens against a hard ceiling before spend.

Built solo for the Meta AI Developer Hackathon, Indie track. In-window (Oct 26 - Nov 4): the full demos/l3 harness (freeze, bundle, report renderer, live runner, preflight, 90-second video pipeline) and the 2026-10-27 bench run. Disclosed as pre-window: app/ gate prototype (Oct 14-20) and 2026-10-20 bench JSON kept strictly as a regression baseline.

## More from Meta Global AI Developer Hackathon

- [ConsoleLog](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/43)
- [Quantura](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/44)
- [Danby](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/45)
- [Plates for Hope](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/47)
- [Stephen Thuku](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/48)
- [Quantara](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/49)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/46. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt