1. Conduid
  2. IoT
  3. Little Canary
MCP server · IoT

Little Canary

by · github.com/hermes-labs-ai/little-canary

Detects prompt injection by its effect on a sacrificial canary model, not just pattern matching: untrusted input hits a powerless model first, a behavioral check reads the residue, and it returns block, flag, or pass before your primary model acts. Inbound preflight sensor, not a guarantee.

Unclaimed agent-securityai-agentsai-safetycanaryclaude-code-plugindefense-in-depthdetectiondeveloper-toolsgemini-cli-extensionhermes-labsinput-validationjailbreak-detectionllm-securitymcppreflightprompt-injectionprompt-injection-detectionpython
InstallView source
39Low

Scored 27 days ago · breakdown

About Little Canary

Little Canary is an MCP server published by hermes-labs-ai in the IoT category: detects prompt injection by its effect on a sacrificial canary model, not just pattern matching: untrusted input hits a powerless model first, a behavioral check reads the residue, and it returns block, flag, or pass before your primary model acts. Inbound preflight sensor, not a guarantee.

The repository has 30 stars and 0 forks, with the last commit 10 days ago. Recent commit activity is the strongest signal that reported issues will get attention.

Install

From source
git clone https://github.com/hermes-labs-ai/little-canary

This server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.

Ask AI

✦

Ask AI about Little Canary

Powered by Claude · Grounded in docs

I know everything about Little Canary. Ask me about installation, configuration, usage, or troubleshooting.

Security checks

6 checks have not been run yet.

Releases

v0.5.0v0.5.0: Ingest surface · 7 Oct 2026Experimental ingest surface with the R1 exact-type/shadowing gate, plus the hardened 0.5.0 screening core. See CHANGELOG.md for details.
v0.4.0little-canary 0.4.0 · 29 Sep 2026Little Canary 0.4.0 Added Offline demo by default: `little-canary demo` runs a packaged replay fixture with zero extra dependencies (no Ollama pull). `--timeout` flag on demo/serve paths; timeout-induced DEGRADED now names the knob.…
v0.3.10little-canary 0.3.10 · 27 Sep 2026Little Canary 0.3.10 This patch publishes the fixes from [PR #112](https://github.com/hermes-labs-ai/little-canary/pull/112) to PyPI. The 0.3.9 package predates that merge. The default canary, detector rules, 256-token output limit, and…
v0.3.9little-canary 0.3.9 · 27 Sep 2026Little Canary 0.3.9 This release makes small local canaries easier to select and evaluate, clarifies incomplete screening, and improves host integration packaging. The default remains `qwen2.5:1.5b`; detection rules and the 256-token…
v0.3.8little-canary 0.3.8 · 18 Sep 2026Packages the Claude Code plugin manifest at the location the Agent Plugins v1.0.0 specification reads, so the plugin resolves for hosts outside Claude Code. No detection, routing, or hook behavior changes. Added…
Questions

About Little Canary

How do I install Little Canary?

No package has been published for this server, so install it from source: clone the repository and follow its README.

Is Little Canary safe to use with an AI agent?

Its trust score is 39 out of 100 (low). Conduid hasn't run static security checks on this repository yet, so review the source yourself before granting it credentials. It has no ConduID identity yet, so agent calls to it are not receipted.

Is Little Canary still maintained?

Yes — the latest release is v0.5.0 (7 Oct 2026), and the last commit was 10 days ago. The repository has 30 stars.