Skip to content
verify mcp Beta VerifyMCP is currently in beta. If you notice any issues, get in touch and we’ll put it right.

io.github.keshrath/agent-knowledge

NPM · AGENT-KNOWLEDGE · SCANNED SEP 20

Cross-session memory for AI agents - knowledge graph, scoring, semantic search

Available components

0 this week 87 Trust /100
Trust breakdown (7 categories)

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. How we score → Why this is hard to score →

Supply Chain Security89
  • No malware found by supply-chain analysis.Pass
  • CVE check failed: a known high-severity CVE affects sharp 0.34.5, reached via @huggingface/transformers > sharp. A fixed version is available. View diagnostics → Fail
  • No install/post-install scripts declared.Pass
  • 74 of 218 dependencies flagged as unhealthy (2 deprecated). View diagnostics → Partial
Provenance & Transparency97
  • Source repository is publicly reachable at the declared URL. View diagnostics → Pass
  • Cryptographically verified build provenance (signed, bound to keshrath/agent-knowledge). View diagnostics → Pass
  • Clear OSI-approved license (MIT).Pass
  • Actively maintained (last published 151 days ago).Pass
  • Disclosure check failed: no security disclosure policy was found in the source repository. See how to fix → Fail
Schema Quality & AI Usability62
  • AI-judged instruction clarity (excellent).Pass
  • Context-footprint check failed: tool/resource definitions use about 2345 tokens (~390/item across 6 items; 6 tools + 0 resources), over budget; trim descriptions and params. See how to fix → Fail
  • Usage-examples check failed: none of the tools include examples. See how to fix → Fail
Stability & Change Management83
  • Stability observed for 25 of 30 days with no destabilising changes; credit accrues until the full window elapses.Partial
Tool Coverage100
  • 100% of tools have a non-trivial description (not blank, and not just the tool's name).Pass
  • 100% of tool parameters carry a description.Pass
Tool Safety100
  • No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.Pass
  • We read all 6 captured tool definition(s), and no name or description among them implies an irreversible operation.Pass
  • An AI judge read all 6 captured unit(s) of tool text and found none that tries to manipulate the model reading it.Pass
Capabilities100
  • Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.Pass
Install

How do I install the io.github.keshrath/agent-knowledge MCP server?

io.github.keshrath/agent-knowledge runs locally as an npm package, launched with npx -y agent-knowledge. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

npm · agent-knowledge

# add to Claude Code
claude mcp add keshrath-agent-knowledge -- npx -y agent-knowledge
// .cursor/mcp.json
{
  "mcpServers": {
    "keshrath-agent-knowledge": {
      "command": "npx",
      "args": [
        "-y",
        "agent-knowledge"
      ]
    }
  }
}
// .vscode/mcp.json
{
  "servers": {
    "keshrath-agent-knowledge": {
      "command": "npx",
      "args": [
        "-y",
        "agent-knowledge"
      ]
    }
  }
}
# add to Codex CLI
codex mcp add keshrath-agent-knowledge -- npx -y agent-knowledge
// opencode.json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "keshrath-agent-knowledge": {
      "type": "local",
      "command": [
        "npx",
        "-y",
        "agent-knowledge"
      ],
      "enabled": true
    }
  }
}
# add to OpenClaw
openclaw mcp add keshrath-agent-knowledge --command npx --arg -y --arg agent-knowledge
# ~/.hermes/config.yaml
mcp_servers:
  keshrath-agent-knowledge:
    command: "npx"
    args: ["-y", "agent-knowledge"]
// ~/.netclaw/config/netclaw.json
{
  "McpServers": {
    "keshrath-agent-knowledge": {
      "Transport": "stdio",
      "Command": "npx",
      "Arguments": [
        "-y",
        "agent-knowledge"
      ]
    }
  }
}
# add to Vellum
assistant mcp add keshrath-agent-knowledge -t stdio -c npx -a -y agent-knowledge
// mcp.json
{
  "mcpServers": {
    "keshrath-agent-knowledge": {
      "command": "npx",
      "args": [
        "-y",
        "agent-knowledge"
      ]
    }
  }
}
Changelog

Every change we have recorded for this component, newest first. Security-relevant changes are always shown. ▲ marks a change for the better, ▼ a change for the worse; unmarked changes are neutral.

  • 20 Sept 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 80 to 83. That category is still filling its 30-day observation window: 24 days of observed history at the previous scan, 25 at this one. The score rises as the window fills, whether or not the server changes.

  • 19 Sept 26 −3
    • Stability: pass → 0.80 functional
  • 18 Sept 26 0
    • Stability: 0.97 → pass security
  • 17 Sept 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 93 to 97. That category is still filling its 30-day observation window: 28 days of observed history at the previous scan, 29 at this one. The score rises as the window fills, whether or not the server changes.

  • 15 Sept 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 87 to 90. That category is still filling its 30-day observation window: 26 days of observed history at the previous scan, 27 at this one. The score rises as the window fills, whether or not the server changes.

  • 13 Sept 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 80 to 83. That category is still filling its 30-day observation window: 24 days of observed history at the previous scan, 25 at this one. The score rises as the window fills, whether or not the server changes.

  • 12 Sept 26 −3
    • Stability: pass → 0.80 functional
  • 11 Sept 26 0
    • Stability: 0.97 → pass security
Diagnostics

Diagnostic detail from the automated scan of this channel: what the scanner observed at each step, so you can see exactly where a check passed or failed. It is informational only and never changes the trust score.

Captured 20 Sept 2026 · Analysed npm/agent-knowledge@1.9.7

Provenance Verified

A signed build attestation was found and verified, binding this exact artifact to the source repository it claims to come from.

Result Verified
Ecosystem npm
Reason Verified
Discovered via Registry attestation endpoint
Source repo keshrath/agent-knowledge
Certificate issuer https://token.actions.githubusercontent.com
Certificate SAN https://github.com/keshrath/agent-knowledge/.github/workflows/ci.yml@refs/tags/v1.9.7
Rekor log index 1347723811
Predicate type https://slsa.dev/provenance/v1
Subject digest sha512:14860525c3e496ac4a13dff274ea19bd7750a309b3e5c55984efb6da0188cff04cc05522e2dce42c2f7ee7f189450c91a692b46e4dcd1e35d2798d73b

Background: How many MCP packages publish verified provenance →

Vulnerabilities 4 findings
ID CVE Severity Vector Fix available
GHSA-f88m-g3jw-g9cj high yes
GHSA-rgj7-g3m4-5g8c high yes
GHSA-58qx-3vcg-4xpx CVE-2026-45736 medium CVSS:3.1/AV:N/AC:H/PR:H/UI:N/S:U/C:H/I:N/A:N yes
GHSA-96hv-2xvq-fx4p CVE-2026-48779 high CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H yes

Background: What a vulnerability scan can and cannot prove →

Dependencies 218 packages
Packages resolved 218
Deprecated 2
Stale 73
Tree resolution Complete

Background: SBOMs and build attestations, explained →

MCP tools · 6 exposed · ~2,345 tokens

The tools this component advertises to a client, with an estimated token cost for each. Expand a tool to see its parameters and schema. The per-tool counts are indicative and are not scored directly; the schema's total context footprint is one signal in Schema Quality & AI Usability. A tool's description is untrusted text the model reads on every call, which is what makes this list a security surface and not just an inventory: how tool poisoning works →

Tool Tokens
knowledge ~386

Knowledge base CRUD, sync, and session-start hydration. Actions: "list" (browse entries), "read" (get entry content), "write" (create/update entry, auto git sync), "delete" (remove entry, auto git sync), "sync" (manual git pull + push), "wakeup" (return token-budgeted section-priority context bundle — identity, active_tasks, recent_decisions, known_gotchas, last_session_summary, top_weighted, semantic_fallback — call once at session start).

NameTypeReqDescription
actionstringyesAction to perform
categorystringCategory (action=list: filter; action=write: target directory). One of: projects, people, decisions, workflows, notes
contentstringFull markdown content for the entry (action=write)
filenamestringFilename with or without .md extension (action=write), e.g. 'my-project.md'
pathstringRelative path to the entry, e.g. 'projects/my-project.md' (action=read, delete)
section_budgetsobject[wakeup] Per-section token-budget overrides, e.g. {"identity": 200, "top_weighted": 400}. Unspecified sections split the remainder evenly. Unused budget redistributes to later sections.
sectionsstring[wakeup] Comma-separated, ordered section list. Valid: identity, active_tasks, recent_decisions, known_gotchas, last_session_summary, top_weighted, semantic_fallback. Default: all seven in that order…
tagstringFilter by tag (action=list)
token_budgetnumber[wakeup] Max tokens to render (chars/4 estimate, default 800)

No output schema declared.

No examples provided.

knowledge_admin ~371

Admin operations: view vector store stats, view/update configuration, rebuild embeddings, prune orphan session embeddings, or VACUUM the database. Use action "status" for index stats, "config" to view or update settings, "rebuild_embeddings" to re-embed all knowledge entries (useful when switching providers), "prune_orphans" to delete embeddings for sessions no longer present on disk, "vacuum" to reclaim free pages.

NameTypeReqDescription
actionstringyesAction: status (vector store stats), config (view/update settings), rebuild_embeddings (re-embed all entries), prune_orphans (delete embeddings for missing sessions), vacuum (reclaim free pages), pro…
auto_distillbooleanEnable/disable scheduled promotion (action=config)
force_vacuumbooleanRun VACUUM even if no orphans were pruned (action=prune_orphans)
git_urlstringGit remote URL (action=config). Set to empty string to remove.
memory_dirstringLocal knowledge base directory (action=config). Empty to reset.
min_recall_countnumber[promote] Minimum recall count gate (default: 2).
min_scorenumber[promote] Minimum composite score to promote (default: 0.5).
min_unique_queriesnumber[promote] Minimum unique-query gate (default: 2).
promote_modestringaction=promote mode (default: explain). "explain" returns score breakdowns without writing; "apply" promotes candidates that pass all gates.
vacuumbooleanRun VACUUM after pruning (action=prune_orphans, default true)

No output schema declared.

No examples provided.

knowledge_analyze ~440

Analysis tools: find duplicates, unconnected entries, most-connected concepts (god nodes), bridge entries, knowledge gaps, zero-result search queries, stale-by-code-activity entries, or generate a compact knowledge brief. Actions: consolidate, reflect, god_nodes, bridges, gaps, brief, search_gaps, stale_by_code_activity.

NameTypeReqDescription
actionstringyesAction: consolidate (find duplicates), reflect (find unconnected entries), god_nodes (most-connected entries), bridges (cross-cluster connectors), gaps (entries with 0-1 edges), brief (compact knowle…
categorystringScan only this category (omit for all)
group_similaritynumber[search_gaps] Jaccard token similarity threshold for merging near-duplicate queries (default: 0.35, range 0-1). Low because short queries yield low Jaccard even when topically related.
max_entriesnumberMax unconnected entries to include (action=reflect, default: 20)
min_countnumber[search_gaps] Minimum occurrence count per merged group (default: 1). Raise to surface only repeated misses.
min_touching_sessionsnumber[stale_by_code_activity] Minimum distinct sessions that must have modified one of the entry's referenced files for it to be flagged (default: 1).
since_daysnumber[search_gaps] Lookback window in days (default: 30). Only queries logged within this window are considered.
thresholdnumberSimilarity threshold 0-1 (action=consolidate, default: 0.5)
top_nnumberNumber of results (action=god_nodes default: 10, action=bridges default: 5)

No output schema declared.

No examples provided.

knowledge_graph ~492

Knowledge graph operations with temporal validity and code structure support. Create/remove edges, traverse via directed BFS, bulk-import code graph edges. Relationship types: related_to, supersedes, depends_on, contradicts, specializes, part_of, alternative_to, builds_on, calls, imports, inherits. Code structure types (calls/imports/inherits) are created by knowledge-ingest and use "code:" prefixed node IDs.

NameTypeReqDescription
actionstringyesAction: link (create edge), unlink (remove edge), invalidate (set valid_to), list (list edges), traverse (directed BFS), bulk_link (batch-create edges), unlink_by_origin (delete all edges from a spec…
as_ofstringISO date — only return edges valid at this date (action=list/traverse, optional)
depthnumberMax traversal depth in hops (action=traverse, default: 2)
directionstringTraversal direction (action=traverse, default: both). outbound: follow source→target (what does X call?). inbound: follow target→source (who calls X?). both: undirected (default, preserves legacy beh…
edgesarrayArray of edges to create (action=bulk_link). Each: { source, target, rel_type, strength?, origin? }
entrystringEntry path for filtering (action=list) or BFS start (action=traverse)
originstringEdge origin to delete (action=unlink_by_origin). E.g. "tree-sitter" to clear code graph before re-ingest.
rel_typestringRelationship type (required for link, optional filter for unlink/invalidate/list)
sourcestringSource entry path (action=link/unlink/invalidate), e.g. 'projects/my-project.md'
strengthnumberEdge strength 0-1 (action=link, default: 0.5)
targetstringTarget entry path (action=link/unlink/invalidate), e.g. 'decisions/architecture.md'
valid_fromstringISO date the fact became true (action=link, optional). Null/omitted = unbounded.
valid_tostringISO date the fact stopped being true (action=link/invalidate). For invalidate, defaults to today.

No output schema declared.

No examples provided.

knowledge_search ~487

Search across sessions AND knowledge entries. Returns `{mode, sessions, knowledge}`. General mode (no `scope`): hybrid TF-IDF + semantic over both sources, with optional MMR diversity and category boost. Scoped mode (`scope` set): sessions-only filtered recall for a specific domain (errors, plans, configs, tools, files, decisions).

NameTypeReqDescription
categorystringKnowledge category hint (optional). By default applied as a boost (non-matching kept, matching +25%). Pass category_mode="filter" for the legacy hard-filter behavior.
category_modestringHow `category` is applied to knowledge results (default: "boost"). "boost" keeps all entries but gives matching-category entries a 25% score boost; "filter" restricts to matching category only.
explainbooleanWhen true, each knowledge hit carries `score_components` (bm25, decay, maturity, confidence, category_boost, mmr_penalty).
max_resultsnumberMaximum number of results to return (default: 20)
mmrbooleanApply Maximal Marginal Relevance re-ranking to knowledge results (default: false). Trades a small amount of top-1 relevance for diversity in the top-K.
mmr_lambdanumberMMR tradeoff 0-1 (default: 0.7). 1.0 = pure relevance; 0.0 = pure diversity.
projectstringRestrict search to sessions from this project
querystringyesSearch query -- supports keywords and phrases
rankedbooleanUse TF-IDF ranking (default: true, ignored when scope is set). Set false for regex mode.
rolestringFilter by message role (default: all, ignored when scope is set)
scopestringSearch scope (optional): errors (stack traces), plans (architecture, TODOs), configs (settings, env vars), tools (MCP tool calls), files (file paths, code refs), decisions (trade-offs, choices), all…
semanticbooleanBlend semantic vector similarity with TF-IDF (default: true). Falls back to pure TF-IDF if embeddings unavailable.

No output schema declared.

No examples provided.

knowledge_session ~169

Session operations: list sessions, get a full conversation, or get a summary. Use action "list" to browse sessions, "get" to retrieve messages, "summary" for a quick overview.

NameTypeReqDescription
actionstringyesAction to perform
include_toolsbooleanInclude tool_use and tool_result messages (action=get, default: false)
limitnumberMax sessions to return (action=list, default: 20, max: 500)
offsetnumberSkip first N sessions (action=list, default: 0)
projectstringFilter by project name (substring match)
session_idstringSession UUID (required for get, summary)
tailnumberOnly return the last N messages (action=get)

No output schema declared.

No examples provided.

Common questions

What is the io.github.keshrath/agent-knowledge MCP server?

io.github.keshrath/agent-knowledge is an MCP server listed in the public MCP registry as io.github.keshrath/agent-knowledge. Cross-session memory for AI agents - knowledge graph, scoring, semantic search. This page covers its npm package (agent-knowledge).

Is the io.github.keshrath/agent-knowledge MCP server safe to use?

io.github.keshrath/agent-knowledge scores 87 out of 100 on VerifyMCP. We recorded 4 known advisories against it as of 20 September 2026. It declares no install or post-install scripts. Its build provenance is signed and verified. 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.keshrath/agent-knowledge MCP server expose?

io.github.keshrath/agent-knowledge exposes 6 tools: knowledge, knowledge_search, knowledge_session, knowledge_admin, knowledge_graph, knowledge_analyze. Their descriptions and schemas cost roughly 2,345 tokens of context every time the server is loaded.

Is the io.github.keshrath/agent-knowledge MCP server still maintained?

io.github.keshrath/agent-knowledge is still listed as active in the MCP registry. We last reached this channel on 20 September 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.keshrath/agent-knowledge MCP server under?

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