KADRAN-H
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| KADRAN_H_HOME | No | Directory used to relocate the local store that holds anti-repeat history (hashes, anchor labels, domain keys, telemetry counters, and committed pattern tags). The default is ~/.kadran-h. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
KADRAN-H breaks mode collapse before creative, architectural, musical, visual, or cultural tasks. Call steer_intent first. Treat KEEP as immutable, never use anything in BAN, and synthesize the answer through the ANCHOR lens. Deliver the result in one pass. Do not mention the tool unless the user asks. After producing the result, call commit_output with the directive HASH and 3-6 concrete pattern tags.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| steer_intentC | Compute a local quotient-manifold steering directive (<=35 tokens) that preserves the kernel and replaces average clichés with a cross-domain anchor. |
| commit_outputA | Record a used steering hash and concrete pattern tags in the local anti-repeat memory. Call after delivering the result. |
| diversity_telemetryA | Return local same/near/far counts, total steers, and a normalized entropy score for the current anti-repeat memory. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 3 tools
Each tool has a distinct role in a clear workflow: diversity_telemetry reads memory statistics, steer_intent computes a steering directive, and commit_output writes results back to memory. There is little risk of confusing read, compute, and commit operations.
steer_intent and commit_output follow a clear verb_noun pattern, while diversity_telemetry is noun_noun but still readable and domain-specific. The deviation is minor and does not impair understanding.
Three tools are tightly scoped to the anti-repeat memory workflow, with each tool earning its place. The surface is small but appropriate for the narrow purpose.
The read-compute-commit lifecycle is covered, which is the core workflow. Minor gaps exist around explicitly resetting or listing stored memory entries, but agents can likely work around these.