# io.github.leocelis/stele (pypi · stele-mcp)

Governed agent memory: lessons stay out of search until another actor promotes them with evidence.

- Trust score: 69/100 (medium)
- Change this week: +7
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-11-21

## Components

- pypi · `stele-mcp`: 69/100 (this document), [markdown](https://verifymcp.io/servers/leocelis-stele/stele-mcp.md), [page](https://verifymcp.io/servers/leocelis-stele/stele-mcp)

## Channel facts

- Registry: `pypi`
- Package: `stele-mcp`
- Version: `18.16.5`
- Transport: `stdio`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically from public evidence about the published package, including repeated runs of it in an isolated sandbox, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-11-21.

- **Supply Chain Security**: 100/100
  - No malware found by supply-chain analysis.
  - No known CVEs affecting this package version or its production dependencies.
  - Runs hatchling.build at install time, a recognised build step with no custom scripting around it.
  - 0 of 29 dependencies flagged as unhealthy.
- **Provenance & Transparency**: 48/100
  - Source repository is publicly reachable at the declared URL.
  - Provenance check failed: no build-provenance attestation is published.
  - Clear OSI-approved license (MIT).
  - Actively maintained (last published 0 days ago).
  - Publishes a security disclosure policy (SECURITY.md).
- **Schema Quality & AI Usability**: 46/100
  - AI-judged instruction clarity (poor).
  - Tool/resource definitions use about 1778 tokens (~49/item across 36 items; 36 tools + 0 resources), lean.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 50/100
  - Stability observed for 15 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 71/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 0% of tool parameters carry a description.
  - Structured output schemas are declared (100% of tools); any adoption earns full credit.
- **Tool Safety**: 75/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - 0 of 2 tool(s) whose name or description implies an irreversible operation declare an MCP destructiveHint annotation; "stele_delete" implies "delete" and declares no destructiveHint at all, which the MCP spec reads as destructive by default.
  - An AI judge read all 36 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.

## Install

### How do I install the io.github.leocelis/stele MCP server?

io.github.leocelis/stele runs locally as a PyPI package, launched with uvx stele-mcp. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

### Claude

```bash
claude mcp add leocelis-stele -- uvx stele-mcp
```

### Cursor

```json
{
  "mcpServers": {
    "leocelis-stele": {
      "command": "uvx",
      "args": [
        "stele-mcp"
      ]
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "leocelis-stele": {
      "command": "uvx",
      "args": [
        "stele-mcp"
      ]
    }
  }
}
```

### Codex

```bash
codex mcp add leocelis-stele -- uvx stele-mcp
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "leocelis-stele": {
      "type": "local",
      "command": [
        "uvx",
        "stele-mcp"
      ],
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add leocelis-stele --command uvx --arg stele-mcp
```

### Hermes

```yaml
mcp_servers:
  leocelis-stele:
    command: "uvx"
    args: ["stele-mcp"]
```

### Netclaw

```json
{
  "McpServers": {
    "leocelis-stele": {
      "Transport": "stdio",
      "Command": "uvx",
      "Arguments": [
        "stele-mcp"
      ]
    }
  }
}
```

### Vellum

```bash
assistant mcp add leocelis-stele -t stdio -c uvx -a stele-mcp
```

### Other

```json
{
  "mcpServers": {
    "leocelis-stele": {
      "command": "uvx",
      "args": [
        "stele-mcp"
      ]
    }
  }
}
```

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-11-21 (score 69, +15)

- [security improvement] Malware scan: unverified → pass

### 2026-11-20 (score 54, +7)

- [security improvement] Malware scan: unverified → pass
- [functional improvement] Stability: unverified → 0.47
- [functional] Package version: 18.16.4 → 18.16.5

### 2026-11-18 (score 47, −15)

- [security regression] Malware scan: pass → unverified

### 2026-11-16 (score 62, +15)

- [security improvement] Malware scan: unverified → pass

### 2026-11-15 (score 47, −15)

- [security regression] Malware scan: pass → unverified

### 2026-11-12 (score 62, +15)

- [security improvement] Malware scan: unverified → pass

### 2026-11-11 (score 47, −15)

- [security regression] Malware scan: pass → unverified

### 2026-11-09 (score 62, +15)

- [security improvement] Malware scan: unverified → pass

## MCP tools (36)

### `stele_add` (~37 tokens)

ADD a distilled structured entry into quarantine. Pass entry as JSON object string.

Input parameters:

- `entry_json` (string, required)
- `ts`

Output parameters:

- `result` (string)

### `stele_add_batch` (~36 tokens)

Atomic multi-ADD under one lock.

Input parameters:

- `actor`
- `entries_json` (string, required)
- `ts`

Output parameters:

- `result` (string)

### `stele_attach` (~60 tokens)

Content-address bytes (base64); optionally LINK to an entry (FF-6).

Input parameters:

- `actor` (string, required)
- `data_base64` (string, required)
- `entry_id`
- `kind` (string)
- `ts`

Output parameters:

- `result` (string)

### `stele_belief_at` (~59 tokens)

Bi-temporal point-in-time belief inventory or SEARCH.

Input parameters:

- `as_of` (string, required)
- `budget` (integer)
- `consumer_scope` (string, required)
- `principal_scopes_json`
- `query`

Output parameters:

- `result` (string)

### `stele_conflict_surface` (~30 tokens)

Conflict-preserving contested pairs — no auto-collapse.

Input parameters:

- `body_max_chars` (integer)

Output parameters:

- `result` (string)

### `stele_delete` (~54 tokens)

DELETE by entry_id or subject_id (true erase + index rebuild).

Input parameters:

- `actor` (string, required)
- `entry_id`
- `reason` (string)
- `subject_id`
- `ts`

Output parameters:

- `result` (string)

### `stele_diff_stores` (~36 tokens)

Diff this store vs another root (e.g. snapshot) by entry id sets.

Input parameters:

- `other_root` (string, required)

Output parameters:

- `result` (string)

### `stele_doctor` (~27 tokens)

Operator health report: verify + stats + contested + stale.

Input parameters:

- `now`

Output parameters:

- `result` (string)

### `stele_entangled_suspects` (~55 tokens)

Human-review queue for LINK-entangled poison (report only).

Input parameters:

- `limit` (integer)
- `seed_ids_json`
- `untrusted_agents_json`
- `untrusted_sources_json`

Output parameters:

- `result` (string)

### `stele_entry_schema` (~24 tokens)

JSON Schema 2020-12 for Stele entries (interop).

Output parameters:

- `result` (string)

### `stele_export` (~73 tokens)

Export a redacted, audience-tiered pack (storage ≠ pack).

Input parameters:

- `audience` (string, required)
- `created_at`
- `dest` (string, required)
- `expiry` (string, required)
- `purpose` (string, required)
- `scope` (string, required)
- `subject_allowlist_json`

Output parameters:

- `result` (string)

### `stele_forget_compliance` (~59 tokens)

Post-erasure active-forgetting probe (GateMem-shaped).

Input parameters:

- `consumer_scope` (string, required)
- `entry_ids_json`
- `forbidden_substrings_json`
- `probe_query`
- `subject_id`

Output parameters:

- `result` (string)

### `stele_hydrate` (~58 tokens)

Import a redacted pack (ADD; optional promote with external evidence).

Input parameters:

- `actor` (string, required)
- `evidence_json`
- `pack_dir` (string, required)
- `promote` (boolean)
- `ts`

Output parameters:

- `result` (string)

### `stele_hygiene_candidates` (~42 tokens)

Zombie / net-harm hygiene report — no auto-delete.

Input parameters:

- `limit` (integer)
- `now`
- `unused_before`

Output parameters:

- `result` (string)

### `stele_lineage` (~31 tokens)

Supersede chain + journal audit lineage (TOKI-shaped).

Input parameters:

- `entry_id` (string, required)

Output parameters:

- `result` (string)

### `stele_link` (~62 tokens)

LINK an entry to an artifact, test, entry, or source.

Input parameters:

- `actor` (string, required)
- `digest`
- `entry_id` (string, required)
- `kind` (string, required)
- `ref` (string, required)
- `ts`

Output parameters:

- `result` (string)

### `stele_list` (~86 tokens)

LIST the calling key's store entries in pages: id, layer, title, state, written_at.

state: "quarantine" | "promoted" | omitted (all live entries). limit: 1-200
(default 50). cursor: the next_cursor from the previous page.

Input parameters:

- `cursor`
- `limit` (integer)
- `state`

Output parameters:

- `result` (string)

### `stele_list_contested` (~23 tokens)

List contested entries awaiting evidenced resolution (TECH_SPEC Q5).

Output parameters:

- `result` (string)

### `stele_pin` (~47 tokens)

Pin or unpin a promoted lesson for SEARCH priority.

Input parameters:

- `actor` (string, required)
- `entry_id` (string, required)
- `pinned` (boolean)
- `ts`

Output parameters:

- `result` (string)

### `stele_promote` (~65 tokens)

PROMOTE a quarantined entry with external-oracle evidence (JSON array).

Input parameters:

- `actor` (string, required)
- `block_injection_suspects` (boolean)
- `entry_id` (string, required)
- `evidence_json` (string, required)
- `ts`

Output parameters:

- `result` (string)

### `stele_purge_by_provenance` (~70 tokens)

Provenance recovery: dry-run or purge entries from untrusted source/agent.

Input parameters:

- `actor` (string, required)
- `dry_run` (boolean)
- `ts`
- `untrusted_agents_json` (string)
- `untrusted_sources_json` (string)

Output parameters:

- `result` (string)

### `stele_record_outcome` (~61 tokens)

Record helpful/harmful/ignored after using a lesson; helpful refreshes last_verified.

Input parameters:

- `actor` (string, required)
- `entry_id` (string, required)
- `note`
- `outcome` (string, required)
- `ts`

Output parameters:

- `result` (string)

### `stele_reflect` (~44 tokens)

REFLECT: dedupe, expire, surface conflicts (no auto-resolve).

Input parameters:

- `actor` (string, required)
- `stale_before`
- `ts`

Output parameters:

- `result` (string)

### `stele_related` (~27 tokens)

Outbound + inbound LINK neighborhood for an entry.

Input parameters:

- `entry_id` (string, required)

Output parameters:

- `result` (string)

### `stele_resolve_contested` (~62 tokens)

Resolve a contested pair by evidenced supersede — never auto-merge.

Input parameters:

- `actor` (string, required)
- `evidence_json` (string, required)
- `loser_id` (string, required)
- `ts`
- `winner_id` (string, required)

Output parameters:

- `result` (string)

### `stele_reverify` (~48 tokens)

Batch-refresh last_verified with external oracle evidence.

Input parameters:

- `actor` (string, required)
- `entry_ids_json` (string, required)
- `evidence_json` (string, required)
- `ts`

Output parameters:

- `result` (string)

### `stele_reviewer_corrections` (~39 tokens)

Bounded contested-first correction slice for reviewer roles.

Input parameters:

- `include_contested` (boolean)
- `limit` (integer)

Output parameters:

- `result` (string)

### `stele_search` (~192 tokens)

SEARCH promoted entries; stale/model policies are flag or withhold.

Input parameters:

- `as_of`
- `body_max_chars`
- `budget` (integer)
- `consumer_domain`
- `consumer_env_json`
- `consumer_model_id`
- `consumer_scope` (string, required)
- `follow_link_depth` (integer)
- `follow_links` (boolean)
- `include_contested` (boolean)
- `min_weibull`
- `model_policy` (string)
- `prefer_fresh` (boolean)
- `prefer_helpful` (boolean)
- `principal_scopes_json`
- `query` (string, required)
- `stale_policy` (string)
- `trusted_sources_json`
- `weibull_eta` (number)
- `weibull_kappa` (number)
- `withhold_injection_suspects` (boolean)

Output parameters:

- `result` (string)

### `stele_snapshot` (~44 tokens)

Cold-copy SoT (manifest, journal, entries, attachments) to dest.

Input parameters:

- `actor` (string, required)
- `dest` (string, required)
- `ts`

Output parameters:

- `result` (string)

### `stele_stale_report` (~26 tokens)

List promoted entries past the staleness horizon.

Input parameters:

- `now`

Output parameters:

- `result` (string)

### `stele_stats` (~26 tokens)

Store health dashboard: counts by state/layer/scope.

Input parameters:

- `now`

Output parameters:

- `result` (string)

### `stele_supersede` (~56 tokens)

SUPERSEDE an old lesson with a new quarantined entry (history kept).

Input parameters:

- `actor` (string, required)
- `new_entry_json` (string, required)
- `old_id` (string, required)
- `ts`

Output parameters:

- `result` (string)

### `stele_timeline` (~24 tokens)

Journal history for one entry.

Input parameters:

- `entry_id` (string, required)

Output parameters:

- `result` (string)

### `stele_update` (~46 tokens)

UPDATE non-state fields on an existing entry.

Input parameters:

- `actor` (string, required)
- `entry_id` (string, required)
- `patch_json` (string, required)
- `ts`

Output parameters:

- `result` (string)

### `stele_verify` (~22 tokens)

C4 integrity report: dual-location, schema, journal parse.

Output parameters:

- `result` (string)

### `stele_verify_pack` (~27 tokens)

Offline pack integrity: stamps + secret scan.

Input parameters:

- `pack_dir` (string, required)

Output parameters:

- `result` (string)

## Diagnostics

Captured diagnostic sections: Provenance, Install scripts, Dependencies. The full working is on the page: https://verifymcp.io/servers/leocelis-stele/stele-mcp#diagnostics

## Score history

- 2026-11-21: 69
- 2026-11-20: 54
- 2026-11-19: 47
- 2026-11-18: 47
- 2026-11-17: 62
- 2026-11-16: 62
- 2026-11-15: 47
- 2026-11-14: 62
- 2026-11-13: 62
- 2026-11-12: 62
- 2026-11-11: 47
- 2026-11-10: 62
- 2026-11-09: 62
- 2026-11-08: 47
- 2026-11-07: 47
- 2026-11-06: 62

## Common questions

### What is the io.github.leocelis/stele MCP server?

io.github.leocelis/stele is an MCP server listed in the public MCP registry as io.github.leocelis/stele. Governed agent memory: lessons stay out of search until another actor promotes them with evidence. This page covers its PyPI package (stele-mcp).

### Is the io.github.leocelis/stele MCP server safe to use?

io.github.leocelis/stele scores 69 out of 100 on VerifyMCP. We found no known CVEs affecting it as of 21 November 2026. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

### What tools does the io.github.leocelis/stele MCP server expose?

io.github.leocelis/stele exposes 36 tools: stele_add, stele_add_batch, stele_attach, stele_belief_at, stele_conflict_surface, and 31 more. Their descriptions and schemas cost roughly 1,778 tokens of context every time the server is loaded.

### Is the io.github.leocelis/stele MCP server still maintained?

io.github.leocelis/stele is still listed as active in the MCP registry. We last reached this channel on 21 November 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

### What licence is the io.github.leocelis/stele MCP server under?

io.github.leocelis/stele declares the MIT licence, which is OSI-approved. That covers the source only, and says nothing about the cost of any service it calls.

### What did VerifyMCP test on the io.github.leocelis/stele MCP server?

VerifyMCP ran 16 checks on io.github.leocelis/stele, covering Malware scan, Known CVEs, Install scripts, Dependency health and 12 more. Endpoint Security and Transport & Reachability do not apply to a PyPI package, so they were not tested. The score does not measure everything, and the scoring documentation lists what it leaves out.

## Links

- PyPI project: https://pypi.org/project/stele-mcp/
- Socket report: https://socket.dev/pypi/package/stele-mcp
- Repository: https://github.com/leocelis/stele
- Changelog RSS feed: https://verifymcp.io/servers/leocelis-stele/stele-mcp.xml
- Changelog JSON feed: https://verifymcp.io/servers/leocelis-stele/stele-mcp/changelog.json
- Trust report (JSON): https://verifymcp.io/servers/leocelis-stele/stele-mcp.json
- HTML version of this page: https://verifymcp.io/servers/leocelis-stele/stele-mcp
