cbrain (chenhong268/cbrain) is an MCP server listed on the M8ven Trust Index. It scores 74 out of 100, grade C. It declares 98 tools. No publisher has claimed this listing.

C
Emerging
74/100

cbrain

Your Agent's Memory, Compounding. Agent 的记忆,复利生长。

Emerging. No concerning findings. Grades remain capped until the project builds reputation through adoption. Grades reflect the full trust pyramid: code, verification depth, and reputation. New projects cap at C until adoption is earned.

How we verified

Code Verified⚡ Live Monitored: not connected

Verified is a snapshot. Live keeps it current, and builds your track record.

⚡ Connect GitHub → continuous verification on every pushwhy connect →

Who stands behind it

chenhong268

Source: github_code

Is this your MCP?

Claim it to get a verified publisher badge, a free copy of our full audit findings, and direct contact for any high-priority issues we find. Or connect your repo for our deepest verification, Live Monitored: read-only, revoke anytime. What we access →

Install from

The grade above is for the source repository. Registries can serve a different version, so we mark the ones we were not able to read.

// key findings
No credential exfiltration, no sensitive file access, no obfuscation
Static analysis found nothing flowing your secrets to unexpected places.
Open source with a license and README
Anyone can audit the code, the license is declared, and the publisher documents what it does.
🔐
You'll be asked for 2 credentials: CBRAIN_CANARY_REGISTRATION_TOKEN, ZHIPU_API_KEY
These are read from process.env at runtime. Make sure you trust where they’ll be sent.
// tools this server exposes98 tools

These names and descriptions are the publisher's own, read from the source code. We print them as written. Our assessment is the findings above, not this list.

run_action_candidates

从 Discovery/Health 信号生成内部行动候选。不发送通知、不创建 issue、不执行修复。

read_action_candidates

读取待处理的内部行动候选。不会执行任何动作。

update_action_candidate_status

更新行动候选状态。支持 resolved/dismissed/seen,不会执行候选动作。

agentic_research
export_grounded_artifact
batch
batch_delete_pages
batch_add_links
batch_merge_pages
brain_storm
get_compounding_reviews
act_on_review_candidate
read_discoveries
run_discovery
find_similar_entities
update_discovery_status

更新发现的处理状态。支持标记已读(seen)、已解决(resolved)、已忽略(dismissed)。

mark_discovery_seen

标记发现为已读。建议使用 update_discovery_status 替代。

dossier
recall_episode
expand_entity
record_feedback
cbrain_recall
graph_query

Query the knowledge graph. Traverse from a seed entity or get backlinks. Accepts a slug or entity name (auto-resolved). Links include source_type (wikilink=human, manual=human explicit input, agent=agent inference, ner=LLM-extracted, dialogue=conversation, writeback=auto) and confidence (0-1, higher

link

Unified link operations. Use action=list/add/remove. Compatibility aliases get_links/add_link/remove_link remain available.

get_links

Get links for a page. Returns outgoing, incoming, or both directions. Links include source_type (wikilink=human, manual=human explicit input, agent=agent inference, ner=LLM-extracted, dialogue=conversation, writeback=auto) and confidence (0-1, higher=more reliable).

add_link

Create a link between two pages. Links created via this tool are marked as source_type=agent (agent_inference) with confidence=0.9. To mark a link as user-confirmed, use confirm_evidence after creation.

remove_link

Remove a link between two pages.

set_hierarchy
get_hierarchy

Get the full hierarchy context for an entity: direct manager, subordinates, and peers.

remove_hierarchy
get_org_tree
ingest

Ingest content into the brain. Supports markdown (with frontmatter) and plain text. IMPORTANT: always provide title and pageType.

ingest_dialogue

Ingest a dialogue/conversation into the brain. Use mode='auto' for automatic background capture (stricter filtering, only high-confidence facts). Use mode='manual' for explicit user-triggered ingestion.

insight
list_insights
get_insight

Get full insight details by ID, including content and linked source entities.

archive_insight

Archive an insight. Archived insights are soft-deleted from active results.

dismiss_insight

Dismiss an insight as not useful. Dismissed insights won't appear in active results.

query_insights
promote_discovery
job

Unified job queue operations. Use action=submit/list/status/cancel/retry. Compatibility aliases remain available.

job_submit

Submit a new job to the queue

job_list

List jobs, optionally filtered by status

job_status

Get detailed status of a specific job

job_cancel

Cancel a pending or running job

job_retry

Retry a failed job

read_knowledge_map
add_knowledge
repair_known_relations

Maintenance-only bounded repair for Markdown Known Relations projection drift. Defaults to dry-run; execute requires an explicit limit.

merge_entities

实体合并专用安全入口。支持 dry_run(返回合并规划,零写入)和 execute(安全执行合并 + 自动验证残留)。仅限 derived 层(entity/concept/insight),拒绝 record 层。返回合并规划(影响分析、冲突检测、wikilink 重写预估)和执行后的验证摘要。

next_actions

只读。返回当前最该处理的 top 1-3 项(跨健康/发现信号)。不修复、不删除、不合并、不提升候选。健康信号来自 run_action_candidates 已分类的候选,不重跑健康检查。

health

Run a 14-dimension health check (errors, dedup, slug collisions, consistency, structural consistency, completeness, islands, suggestions, attention, data readiness, source quality, etc.). Returns issues and writes a report file.

enrich

Run entity enrichment. Upgrades entity tiers based on mention counts.

writeback

Write insights back to the knowledge base. Actions: 'append' (add content to existing page), 'create_concept' (create new concept page), 'create_link' (add relation between two pages). All operations are logged.

generate_indexes

Generate Obsidian-readable index files: All-Entities, All-Concepts, All-Sources, Dashboard.

status

Get brain status: page counts, sync info, watcher state, quarantine, etc.

get_ingest_log

Get recent ingest log entries.

dream

异步执行完整夜间维护流程(sync → enrich → seal → cleanup → health → insight archive)。立即返回 job_id,后台执行。使用 dream_status 查询进度。有循环锁防止重叠执行。

dream_status

查询最近一次 dream 任务的状态和阶段进度。

dream_reset

Clear the dream cycle lock. Use when a previous dream didn't finish and you need to force a new one.

watcher_quarantine

Manage watcher quarantine and bulk-change backpressure. 'list' shows quarantined files. 'release'/'release_all' manages quarantine. 'bulk_status' shows bulk-pending state. 'bulk_resume' resumes paused bulk processing.

relation_audit

Audit and fix non-standard relation types. Modes: 'report' shows distribution, 'fix' migrates non-standard relations to canonical types. Use dry_run=true to preview changes before applying.

get_page

Get a page by slug. Returns frontmatter + body.

list_pages

List pages in the brain. Optional type filter.

put_page
append_page

Append content to an existing page's body. Does NOT overwrite existing content. Triggers re-index and NER/wikilink extraction.

delete_page

Delete a page by slug. Removes both the vault file and database entry.

resolve_slugs

Resolve page titles or partial names to slugs. Returns best match for each query.

merge_pages

Merge a source page into a target page. All links, timeline entries, tags and raw data are moved from source to target. Source body is appended to target body. Source page is deleted after merge. Use dryRun=true to preview without executing.

alias

Unified alias operations. Use action=add/remove. Compatibility aliases add_alias/remove_alias remain available.

add_alias

Add an alias to a page. NER will resolve the alias to this page instead of creating a new entity.

remove_alias

Remove an alias from a page.

get_pages
profile
get_profile
update_profile
remove_profile
reload_profile
read_project_state

读取 compact 项目状态 artifact,用于 Agent session continuity。只读,不写 vault/SQLite/LanceDB,不做 prompt 注入。

get_provenance

获取知识条目的溯源信息:来源、信任状态、纠正历史

set_trust_state

设置知识条目的信任状态(仅降级/纠正,不可升级为 trusted)。要将条目升级为 trusted,请使用 confirm_evidence。

confirm_evidence

用户明确确认一条知识为可信事实。必须提供 confirmation_record_slug(vault 中已存在的页面)和 excerpt(该页面正文中必须包含的确认原文),系统会验证页面存在且 excerpt 出现在页面正文中。仅当用户在对话中明确表示认可时使用。

deep_recall
query
get_chunks

Get indexed text chunks for a page.

summarize
sync

Sync vault files to SQLite + LanceDB indexes.

remove_orphans

Remove database entries that have no corresponding vault file. Run this after manually deleting files from the vault.

tag

Unified tag operations. Use action=list/add/remove. Compatibility aliases get_tags/add_tag/remove_tag remain available.

get_tags

Get all tags for a page.

add_tag

Add a tag to a page. Updates both the database and the vault file frontmatter.

remove_tag

Remove a tag from a page. Updates both the database and the vault file frontmatter.

timeline

Unified timeline operations. Use action=get/add. Compatibility aliases get_timeline/add_timeline_entry remain available.

get_timeline

Get timeline entries for a page.

add_timeline_entry

Add a timeline entry to a page.

get_versions

Get version history for a page.

revert_version

Revert a page to a specific version. Creates a version snapshot before reverting.

wakeup_diff

生成认知变化摘要(Wake-up Diff)。对比上次快照,产出新增记忆项、内容更新、tier 变化、关系变化、置信度衰减等差异。首次运行建立基线。可由 dream 自动触发或手动运行。

// environment variables
To run this server yourself, you supply these values. They go in your own MCP client configuration and stay on your machine. The secret label means the value is sensitive, not that the server mishandles it.
configCBRAIN_CANARY_FAULT
configCBRAIN_CANARY_PARENT_MANAGED_DIR
configCBRAIN_CANARY_PROCESS_REGISTRY
🔐 secretCBRAIN_CANARY_REGISTRATION_TOKEN
configCBRAIN_CANARY_SNAPSHOT_ROOT
configCBRAIN_CANARY_WRAPPER_IDENTITY
configCBRAIN_CONFIG
configCBRAIN_INGEST_NER_MODE
configCBRAIN_LOCK_ID
configCBRAIN_MCP_TOOL_PROFILEWhole-runtime selection (stdio / env): =agentmaintenancedebugfull (default full; env only, no config field yet). Invalid values fail fast at startup.
configCBRAIN_OUTPUT_BOUNDARY
configCBRAIN_REQUIRED_SKILL_TARGETS
configCBRAIN_ROLLBACK_CANDIDATE
configCBRAIN_ROLLOUT_COHORT_ID
configCBRAIN_ROLLOUT_CONFIG_IDENTITY
configCBRAIN_ROLLOUT_DEPLOYMENT_DIGEST
configCBRAIN_UNSAFE_ALLOW_MULTI_WRITER
configCBRAIN_VAULT
configDOCS_DIR
configGATE_PROJECT_DIR
configHERMES_FAULT_HANG
configHERMES_FAULT_PRIVACY_LEAK
configHERMES_FAULT_QUERY_BUDGET
configHERMES_FAULT_RETRIEVAL
configRC_FAULT_HANG
configRC_FAULT_PRIVACY_LEAK
configRC_FAULT_QUERY_BUDGET
configRC_FAULT_RETRIEVAL
configSKILLS_INDEX
configXDG_CONFIG_HOME
configXDG_DATA_HOME
🔐 secretZHIPU_API_KEY
configHERMES_MANAGED_DIR
// quality suggestions

Tool annotations

No tools have read-only/destructive annotations

Add readOnlyHint or destructiveHint annotations to every tool so hosts can warn users before invoking.

All four hints declared on every tool

98/98 tools missing one or more hints — run_action_candidates (missing: readOnlyHint, destructiveHint, idempotentHint, openWorldHint); read_action_candidates (missing: readOnlyHint, destructiveHint, idempotentHint, openWorldHint); update_action_candidate_status (missing: readOnlyHint, destructiveHint, idempotentHint, openWorldHint), +95 more. OpenAI's directory rejects tools where any of the four hints are missing or non-boolean.

For every tool, set all four hints (readOnlyHint, destructiveHint, idempotentHint, openWorldHint) to explicit true/false values that match the handler’s actual behaviour.

Shell command execution

7 calls in production code run through a shell (bin/check-first-recall-gate.ts:290, bin/check-first-recall-gate.ts:303, bin/check-install-network.ts:60)

Prefer library functions over shell-outs. If you must shell out, ensure all inputs are properly escaped.

Secrets not logged

57 secret values sent to error/log

Redact or omit secret values from log output.

Dependency freshness

1/8 production deps stale: gray-matter@2023-07-12 (3.1y)

Claim the listing to review these findings one by one and send us a correction where you disagree, straight to the team. Claiming also means we tell you when the grade moves, and reach you first if we find anything urgent.

// full audit trail
The findings above are the summary. The full trail, every check we ran, each deduction, the network hosts observed and the dependency advisories, goes to verified publishers, along with an alert whenever a new one lands. Verified publishers can also review each finding and dispute it in one click. Publisher corrections have sharpened several of our checks this month, because the maintainer knows the codebase better than any scanner.
// improvement guidance — verified publishers only
We have 4 concrete improvements we can share with the publisher of this MCP. Each comes with specific guidance to raise the trust score.
// embed badge in your README
[![M8ven Score](https://m8ven.ai/badge/mcp/chenhong268/cbrain)](https://m8ven.ai/mcp/chenhong268/cbrain)
Shows your grade and updates automatically. Prefer no grade? Append ?variant=verified to the badge URL.
commit: e5f8fbe5f88628dad531df8a7902692f219939ff
code hash: 0ec1dd27d49876d6cd700923c408f8c0ab7711d1738c30fe914ab07c3a8c4736
view raw JSON →
Check MCPs from inside your assistant
Tool Check · MCP

Vetting this one by hand? Tool Check is an MCP that scores other MCPs. Add it once and ask Claude, ChatGPT, or any MCP client to grade a server, surface CVEs, check the publisher, and suggest safer alternatives — before you install.

https://m8ven.ai/api/mcp/tool-check
check_toolsearch_toolscompare_toolsrecommend_alternativescheck_publisherreport_concern
How to add it →Free · no account needed · works in any MCP client