Centaur
Explore
Submit
Forum
About
|
Account
Forum
/
Show & Tell
Show & Tell
Share what you have built
0
Witness — Human-in-the-loop coordination primitive. Surfaces decisions requiring human judgment, blocks until approval/rejection, captures rationale, writes audit.
jmcentire
·
May 7
·
0 replies
0
Vigil — Sub-threshold anomaly detection. Watches event streams, trace samples, and contract violations for pattern divergence that doesn't trip any single alarm.
jmcentire
·
May 7
·
0 replies
0
Scram — Emergency kill-switch service. The exemplar stack's last-resort termination mechanism. Does not negotiate, does not retry — terminates and rolls back.
jmcentire
·
May 7
·
0 replies
0
Covenant — Contract validation runtime. Per-environment violation policy + sliding-window violation budget. Twin TypeScript + Python siblings.
jmcentire
·
May 7
·
0 replies
0
Aegis — Resource-budget primitive. Wrap every blocking call with a declared timeout, optional fallback, and structured tags. Twin TypeScript + Python implementations.
jmcentire
·
May 7
·
0 replies
0
BlindDB — Interactive demo of client-side encryption architecture. Server stores only encrypted blobs at opaque addresses; credentials never leave the client.
jmcentire
·
May 7
·
0 replies
0
HermesP2P — Decentralized, ephemeral peer-to-peer messaging. No servers, no persistence, no compromise. PGP-style encryption, public/private channels.
jmcentire
·
May 7
·
0 replies
0
Tessera — Self-validating executable document format. The document proves itself; the engine validates the chain; nothing outside the artifact is consulted.
jmcentire
·
May 7
·
0 replies
0
Ascend — AI-powered engineering management CLI. Roster, meeting transcripts, risk detection, coaching, OKR/PIP tracking, interactive TUI.
jmcentire
·
May 7
·
0 replies
0
Agent-Safe — Embed authorization in the token. 150 lines. Zero deps. Microseconds. SPL (Safe Policy Lisp) inside signed capability tokens.
jmcentire
·
May 7
·
0 replies
0
Signet-Eval — Deterministic policy enforcement for AI agent tool calls. No LLM in the authorization path; 25ms end-to-end.
jmcentire
·
May 7
·
0 replies
0
Signet-Eval — Deterministic policy enforcement for AI agent tool calls. No LLM in the authorization path; 25ms end-to-end.
jmcentire
·
May 7
·
0 replies
0
ZeroGPT Plus — ZeroGPT Plus is a free, no-login AI content detector and humanizer that identifies AI-generated text. It offers multi-language support to refine AI content into natural-sounding, human-like prose.
coupongpts-eng
·
Apr 3
·
1 replies
1
Herald — Async webhook relay and message queue for AI agents. Pull-based, encrypted, self-hostable.
jmcentire
·
Mar 30
·
0 replies
1
Advocate — Six-persona adversarial review engine
jmcentire
·
Mar 30
·
0 replies
1
Transmogrifier — Register-aware prompt translation. Detects linguistic register (casual, technical, academic, narrative, direct) and normalizes it to maximize LLM output quality. Empirically validated: 18.8pp spread o
jmcentire
·
Mar 30
·
0 replies
1
Webprobe — Generic site state-graph auditor with LLM-driven exploration
jmcentire
·
Mar 30
·
0 replies
1
Kindex — Persistent knowledge graph for AI coding assistants
jmcentire
·
Mar 25
·
0 replies
1
Exemplar — Governed code review service demonstrating the full governance stack.
jmcentire
·
Mar 25
·
0 replies
1
Cartographer — Stack adoption and compatibility scanning for existing codebases.
jmcentire
·
Mar 25
·
0 replies