memory-seed (jnl-tshi/memory-seed) is an MCP server listed on the M8ven Trust Index. It scores 74 out of 100, grade C. It declares 32 tools. No publisher has claimed this listing.

C
Warning
74/100

memory-seed

Portable local memory system and MCP search tooling for file-reading AI coding agents.

Warning. Serious findings were identified. Review the full report before connecting. Grades reflect the full trust pyramid: code, verification depth, and reputation. New projects cap at C until adoption is earned.

How we 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

jnl-tshi

Source: github_repo_search

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
🚨
Code appears obfuscated
11 files are unreadable to a human reviewer. Cannot audit what they do.
// tools this server exposes32 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.

memory_decision_provenance

Show a decision's validated Git-reference bindings and temporary before/after code projections. Read-only; code is unavailable rather than inferred when Git evidence cannot be resolved.

memory_decision_provenance_bind

Validate and append one reference-only provenance binding to its owning runtime sidecar. Default is a dry run; set apply true to write. Uses the same validation and output shape as CLI provenance bind.

memory_retrieval_spec_preview

Validate and plan one inline Retrieval Specification or exact local profile against canonical Markdown. Read-only; creates no Evidence Pack.

memory_retrieval_spec_resolve

Resolve one inline Retrieval Specification or exact local profile into an ephemeral Evidence Pack returned inline. Read-only; no cache, registry, provider, or Trace dependency.

memory_task_packet_preview

Validate, measure, resolve, and return one complete deterministic Task Packet inline. Read-only; it does not export, dispatch workers, or create worktrees.

memory_task_packet_compile

Compile and return one complete deterministic Task Packet inline. Read-only; MCP never exports a file or dispatches workers.

memory_search

Search local Memory Seed session logs and return ranked, source-linked context chunks.

memory_link_suggest

Rank older session entries to link from a target entry, closing the authoring loop: returns paste-ready related_entries candidates. Read-only; the agent writes the edge into its own new entry.

memory_link_show

Show one entry's related-entry graph node: stored outbound edges, computed inbound backlinks, supersession edges, importance score, and linked-commit count. Read-only graph traversal.

memory_links_chain

Return the derived refines lifecycle chain for one known entry or decision ref: root, ordered members, head, length, and owning ADR evidence. Read-only.

memory_link_audit

Return judgment-ready lifecycle gap candidates and ranking provenance. Mirrors the read-only part of `link audit`; it cannot apply or scaffold sidecars.

memory_esr

Return the complete structured End-of-Session Report. Equivalent to `esr --json`; it never repairs authoritative memory, but may incrementally refresh the rebuildable ignored temporal-lineage cache and other derived cache state.

memory_decision_provenance_check

Audit one runtime-owned provenance sidecar and its Git reference evidence. Read-only; unavailable Git is reported as unverifiable rather than raised.

memory_branch_status

Read Git branch/worktree posture and return Memory Seed branch-history guidance.

memory_worktree_guard

Classify whether the current Git worktree is safe for a named agent. Read-only pre-write guard.

memory_session_fuse_preview

Dry-run branch-local session entry and diagram-sidecar fuse planning. Read-only; use the CLI --apply path during an in-progress merge to write.

memory_get_chunk

Fetch an exact Memory Seed chunk by chunk_id. Decision ids (entry_id:d1) resolve, and a decision payload carries `entry_context`: the entry-level sections that frame it - Summary, Follow-up, Validation - since a decision block alone omits what the session was doing and what was left open. Sibling de

memory_topics_list

List the controlled topic vocabulary from .memory-seed/topics.yaml. Read-only.

memory_topic_inspect

Inspect one controlled topic or alias and show matching session entries. Read-only.

memory_topics_check

Validate controlled topic vocabulary shape and entry topic usage. Mirrors `memory-seed topics check`; read-only.

memory_adrs_list

List living architectural decision threads and their replay-derived authoritative heads. Read-only.

memory_adr_show

Read one composed ADR sidecar and return its replay-derived status and source decision.

memory_adr_review

Resolve lifecycle decision references to every ADR whose current or historical lineage they affect. Read-only.

memory_adr_reviewed

Append a reviewed-no-change event to one ADR using an existing session entry as the evidence anchor. This moves no ADR head and creates no session entry. CLI twin: memory-seed adr reviewed --adr-id <id> --entry <entry-id> --reason <text>.

memory_adrs_check

Validate every ADR sidecar and replay its append-only lifecycle. Read-only.

memory_session_append

Append a session entry with every structural guarantee enforced. THIS IS THE ONLY WAY TO AUTHOR AN ENTRY - do not hand-write session files. The tool owns structure (target path, heading timestamp from the server

memory_link_retract

Retract a published lifecycle edge, append-only, and optionally re-author it under the right kind or evolution type in the same block. THIS IS THE MANDATED FIX for the `links check` errors `untyped-evolves`, `unknown-evolution-type` and `multiple-refines-successors`: the declaring block is never reo

memory_session_integrate

Merge a task branch and fuse its branch-local session memory in one step, so branch entries land in chronological order on the trunk instead of wherever a raw line-merge puts them. Fails closed: any session

memory_reflection_board_view

Read every active Reflection Board v1 candidate from the current branch's trusted Git history. Read-only; malformed and unsupported reserved candidates remain visible.

memory_reflection_ledger_view

Read one committed, history-classified Reflection Board v1 ledger. Read-only; this never reads an uncommitted ledger suffix.

memory_reflection_ledger_init

Preview or apply initialization of the current branch's only Reflection Board v1 ledger through the kernel-owned transaction writer. The server never accepts a caller-supplied ID, ledger bytes, Git commit, or ref.

memory_reflection_ledger_append

Preview or apply one Reflection Board v1 record through the same kernel-owned transaction writer as the CLI. It never accepts raw ledger bytes, Git identities, receipts, or a history override.

// 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.
configCODEX_SANDBOX
configMINIFLARE_REGISTRY_PATH
configWRANGLER_LOG_PATH
configWRANGLER_WRITE_LOGS
configPYTHONHASHSEED
configPYTHONPATH, reuses a healthy Trace server already on its requested port, and opens the React UI at /.
configMEMORY_SEED_USER
configMEMORY_SEED_CORPUS_CACHE_DIR
configXDG_CACHE_HOME
configMEMORY_SEED_CACHE_DIR
configMEMORY_SEED_LENSE_CACHE_ROOT
configMEMORY_TRACE_STATIC_ROOT
configMEMORY_SEED_LENSE_SKIP_BROWSER
configCUDA_VISIBLE_DEVICES
// 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

32/32 tools missing one or more hints — memory_decision_provenance (missing: readOnlyHint, destructiveHint, idempotentHint, openWorldHint); memory_decision_provenance_bind (missing: readOnlyHint, destructiveHint, idempotentHint, openWorldHint); memory_retrieval_spec_preview (missing: readOnlyHint, destructiveHint, idempotentHint, openWorldHint), +29 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.

Readable source code

11 files are minified or bundled, which is usually build output rather than concealment

Ship unminified, readable source.

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 3 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 Verified](https://m8ven.ai/badge/mcp/jnl-tshi/memory-seed?variant=verified)](https://m8ven.ai/mcp/jnl-tshi/memory-seed)
Shows verification status without the grade. Want the grade badge instead? Remove ?variant=verified from the URL.
commit: d781a271fe3ffb810b1a53b8a860dc5ceb48ac39
code hash: 86f80d24c28f3be0276dc77240ff869d52f07e86dee8c5cc0779f2aa89e57253
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