llm-kosh
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| LLMKOSH_ROOT | No | The root directory for the cartridge (alternative to --root) |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| search_memoryC | Search the Cartridge Knowledge Base for memories, decisions, and files. |
| get_cartridge_memory_mapB | Returns a structural map of the knowledge base projects and active concepts. |
| get_project_contextC | Fetches all context and active decisions for a specific project. |
| list_intakeB | Lists the current items pending in the intake queue. |
| get_daemon_statusC | Verifies the cryptographic ledger of the Cartridge. |
| company_memory_searchC | Permission-first hybrid search over atomic, evidence-backed company memory. |
| company_context_compileD | Compile structured, cited, token-budgeted context for an agent task. |
| company_memory_getB | Get one authorized memory with its evidence references and lifecycle. |
| company_brain_healthB | Validate canonical company-brain storage and projection consistency. |
| company_brain_evaluateB | Run reference storage, citation, and projection acceptance checks. |
| company_session_understandC | Build normalized sessions, episodes, and cited candidates from JSONL evidence. |
| company_sessions_listC | List authorized normalized source sessions. |
| company_episodes_searchC | Search authorized goal-oriented work episodes and observed outcomes. |
| company_episode_getB | Get one authorized episode with ordered event and candidate provenance. |
| company_artifact_registerB | Register an existing local artifact by fingerprint without copying its bytes. |
| company_artifact_inspectC | Verify and inspect a bounded region of an authorized registered artifact. |
| company_artifact_segmentC | Inspect an artifact and persist bounded derived segments with native citations. |
| company_artifact_snapshotC | Explicitly materialize an authorized artifact into immutable snapshot storage. |
| company_memory_proposeC | Create immutable evidence and a non-authoritative candidate memory. |
| company_memory_propose_from_evidenceC | Create a candidate semantic memory citing an existing registered artifact. |
| company_memory_reviewC | Apply a governed lifecycle transition to an authorized memory. |
| create_private_context_packC | Creates a highly focused context pack for a specific task. Includes private and sensitive information. |
| submit_memory_receiptA | Submits a MEMORY_RECEIPT.md formatted string to the intake queue. Does not apply it directly; it must be reviewed and applied. |
| intake_convert_fileC | Converts a local file (e.g. PDF, DOCX, XLSX, PPTX, PNG, WAV) to structured markdown using the MarkItDown converter and ingests it into the cartridge. |
| apply_intake_proposalC | Directly applies an intake proposal batch to the live memory. |
| reasoning_ingestA | Add a bounded atomic fact to the Temporal Causal Reasoning Graph. This is a write operation and requires --allow-write. documented_at and valid_from are ISO 8601 datetime strings. causal_edges: JSON array of {"target_id": str, "edge_type": str, "confidence": float}. Returns the new fact_id. |
| reasoning_queryB | Query the Temporal Causal Reasoning Graph. By default returns a human-readable causal narrative (narrative=True). Set narrative=False to get raw JSON with full fiber bundle details. temporal_context: ISO 8601 datetime or Unix timestamp string (omit for now). |
| reasoning_critiqueA | Run the Lyapunov critic on a specific list of fact IDs. Returns stability score, status, and per-dimension breakdown. |
| reasoning_exploreB | Enumerate all causal paths between two known facts. Returns the fiber bundle for that specific pair. |
| kosh_verifyC | Verify a question against local temporal/causal evidence. This is read-only. The returned JSON report can include supporting facts, causal paths, contradictions, inferred-but-not-discovered relationships, missing evidence, stability information, and an explicit abstention state. It does not consult the internet, seed demo data, write or mutate memory, export private context, or make an imported source trustworthy. |
| trusted_memory_recallC | Recall governed memory with admission and lifecycle policy enforced. |
| trusted_memory_inboxB | List authorized candidate and quarantined Trusted Memory items. |
| trusted_memory_conflictsC | List unresolved authorized Trusted Memory conflicts. |
| trusted_memory_explainB | Explain one authorized memory with evidence and admission history. |
| trusted_memory_proposeA | Propose governed memory. Requires MCP write capability. If |
| trusted_memory_reviewC | Explicitly approve, quarantine, reject, or supersede governed memory. Requires the existing MCP mutate capability. Conflict resolution is never inferred from a generic approval. |
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 36 tools
There are multiple overlapping retrieval and governance clusters: search_memory, company_memory_search, trusted_memory_recall, and kosh_verify all serve different-but-similar memory/questioning purposes, while company_brain_health and company_brain_evaluate appear nearly interchangeable. The trusted_memory_* and company_memory_* families further blur boundaries around 'governed memory' vs 'company memory'.
Naming mixes several patterns: verb-first (search_memory, list_intake), noun-first domain actions (trusted_memory_propose, company_memory_search), and prefix-plus-verb (reasoning_query, intake_convert_file). Although all names are snake_case, the inconsistent ordering and domain-prefix placement make the surface harder to predict.
36 tools is heavy for a single MCP server, spanning at least four major subsystems: cartridge/intake, trusted memory, company memory/artifacts, and causal reasoning. Many tools could reasonably be split into separate servers, and the count creates cognitive load and selection risk.
The server covers many lifecycle stages across memory, evidence, artifacts, intake, and reasoning, including propose/review/recall/search and artifact registration/inspection/snapshot. However, there are notable gaps such as no direct memory deletion/update, no artifact listing, and no project listing beyond a structural map.