MCP server · Org-owned · Audited from source

samyama-graph MCP server

by samyama-ai/samyama-graphApache-2.0 licenceAudited 6 Oct 2026 · source dca807886561

npm samyama-sdk 1.11.0★ 176

✕BLOCKGrade D · 63/100 · 6 Oct 2026 · dca8078

Safer alternatives View source on GitHub ↗

Is it safe? No — BLOCK, grade D (63/100). It exposes 12 tools: 10 read, 2 write and 0 destructive. The audit recorded 22 findings (0 critical, 3 high, 1 medium, 13 low). Maintainer's description: Rust graph-vector database: OpenCypher (99.9% of evaluated TCK scenarios pass), vector search, graph algorithms, RESP + HTTP. LDBC SNB Interactive and BI complete; 1B edges on one machine.

Ask an AI about this pageGChatGPT CClaude PPerplexity GeGemini XGrok

CoCopilot AIGoogle AI Mode

Cite

Cite this page

ModelsAtlas. “samyama-graph MCP server (samyama-ai/samyama-graph)”. Retrieved 6 Oct 2026. https://modelsatlas.com/mcp/samyama-ai/samyama-graph

Why this is BLOCK. The audit found 0 critical and 3 high findings; the most severe is Code injection at src/vector/index.rs:102. Each finding is listed below with its file and line.

Why this page is still public. People search for this server by name. Hiding it would leave them with the README's word alone; here they see what the code does. We list it so you can decide, and we do not show connect commands for BLOCK servers.

Trust audit

Audit 0.4.1 · full source read 6 Oct 2026

Trust score63/100✕BLOCK D

0 critical3 high1 medium13 low5 info

22 findings, listed below with the file and line each was found on. Grade A needs a sandboxed behavioural run too; this audit read the source only.

What the source does

FilesystemDeclared4 observations
NetworkDeclared5 observations
ShellDeclared3 observations
DependenciesNot pinnedSome ranges float
SecretsNone foundNo keys in source

Findings (22)

Sorted by severity

HIGHCode injectioncode.eval_exec · src/vector/index.rs:102

F-02 Code injectionCWE-78 · CWE-94 · CWE-95Layer L1

fn eval(&self, va: &[f32], vb: &[f32]) -> f32 {

Why: evaluates text as code
Fix: remove; use a parser or a dispatch table

HIGHCode injectioncode.eval_exec · src/vector/index.rs:146

F-02 Code injectionCWE-78 · CWE-94 · CWE-95Layer L1

fn eval(&self, va: &[f16], vb: &[f16]) -> f32 {

Why: evaluates text as code
Fix: remove; use a parser or a dispatch table

HIGHCode injectioncode.eval_exec · src/vector/index.rs:209

F-02 Code injectionCWE-78 · CWE-94 · CWE-95Layer L1

fn eval(&self, va: &[f32], vb: &[f32]) -> f32 {

Why: evaluates text as code
Fix: remove; use a parser or a dispatch table

MEDIUMNetwork egressnet.raw_ip · case_studies/_lib/demo_lib.py:88

F-04 Network egressCWE-200 · CWE-319Layer L1

base_url = base_url or os.environ.get("SG_BASE_URL", "http://127.0.0.1:8080")

Why: No further detail recorded for this rule.
Fix: —

LOWCode injectioncode.eval_exec · examples/ic1_index_regression.rs:27

F-02 Code injectionCWE-78 · CWE-94 · CWE-95Layer L1

fn exec(s: &GraphStore, q: &str) -> (usize, Duration) {

Why: evaluates text as code
Fix: remove; use a parser or a dispatch table

LOWCode injectioncode.eval_exec · examples/ic1_index_regression.rs:53

F-02 Code injectionCWE-78 · CWE-94 · CWE-95Layer L1

let (r, d) = exec(s, q);

Why: evaluates text as code
Fix: remove; use a parser or a dispatch table

LOWFilesystem / pathfs.traversal · src/query/csv_source.rs:177

F-05 Filesystem / pathCWE-22 · CWE-59Layer L1

resolve("../../etc/passwd"),

Why: No further detail recorded for this rule.
Fix: —

LOWNetwork egressnet.beacon_words · examples/enterprise_soc_demo.rs:470

F-04 Network egressCWE-200 · CWE-319Layer L1

MitreDef { technique_id: "T1048",     name: "Exfiltration Over Alternative Protocol", tactic: "Exfiltration" },

Why: No further detail recorded for this rule.
Fix: —

LOWNetwork egressnet.beacon_words · examples/enterprise_soc_demo.rs:623

F-04 Network egressCWE-200 · CWE-319Layer L1

AttackStep { step: 5, source_name: "int-app-01",   target_name: "int-db-01",      technique: "T1048",     description: "Database credentials exfiltrated via DNS tunneling" },

Why: No further detail recorded for this rule.
Fix: —

LOWNetwork egressnet.beacon_words · examples/enterprise_soc_demo.rs:730

F-04 Network egressCWE-200 · CWE-319Layer L1

println!("    Incoming IOC: Suspicious PowerShell beacon pattern detected.");

Why: No further detail recorded for this rule.
Fix: —

LOWNetwork egressnet.raw_ip · .github/workflows/tag-release.yml:230

F-04 Network egressCWE-200 · CWE-319Layer L1

curl -sf http://127.0.0.1:8080/api/status >/dev/null && break

Why: No further detail recorded for this rule.
Fix: —

LOWNetwork egressnet.raw_ip · .github/workflows/tag-release.yml:233

F-04 Network egressCWE-200 · CWE-319Layer L1

curl -sf http://127.0.0.1:8080/api/status

Why: No further detail recorded for this rule.
Fix: —

LOWNetwork egressnet.raw_ip · .github/workflows/tag-release.yml:234

F-04 Network egressCWE-200 · CWE-319Layer L1

curl -sf -X POST http://127.0.0.1:8080/api/query \

Why: No further detail recorded for this rule.
Fix: —

LOWNetwork egressnet.raw_ip · .github/workflows/tag-release.yml:237

F-04 Network egressCWE-200 · CWE-319Layer L1

curl -sf -X POST http://127.0.0.1:8080/api/query \

Why: No further detail recorded for this rule.
Fix: —

LOWObfuscation / stealthobf.homoglyph · examples/uc3_capacity_planning.rs:277

F-07 Obfuscation / stealthCWE-506 · CWE-94Layer L1

"   {fid}: Δbeds={:+.1} Δnurses={:+.1} Δdoctors={:+.1}",

Why: No further detail recorded for this rule.
Fix: —

LOWSupply chainsupply.unpinned · sdk/python/examples/mcp-demo/requirements.txt

F-08 Supply chainCWE-829 · CWE-1357Layer L1

samyama, PyYAML

Why: 2 requirement(s) not pinned with ==
Fix: pin exact versions

LOWSupply chainsupply.unpinned · sdk/typescript/package.json

F-08 Supply chainCWE-829 · CWE-1357Layer L1

@types/node, typescript

Why: 2 dependency range(s) float
Fix: pin exact versions or ship a lockfile

INFOInventory / provenanceinv.oversize · case_studies/legal-judgments/cloud-demo.gif

F-13 Inventory / provenanceCWE-1104Layer L0

case_studies/legal-judgments/cloud-demo.gif

Why: 5609783 bytes not read
Fix: —

INFOInventory / provenanceinv.oversize · ldbc-benchmark-results.png

F-13 Inventory / provenanceCWE-1104Layer L0

ldbc-benchmark-results.png

Why: 1410926 bytes not read
Fix: —

INFOInventory / provenanceinv.oversize · sdk/python/examples/mcp-demo/samyama-mcp-demo.gif

F-13 Inventory / provenanceCWE-1104Layer L0

sdk/python/examples/mcp-demo/samyama-mcp-demo.gif

Why: 1178538 bytes not read
Fix: —

INFOInventory / provenanceinv.oversize · sdk/python/examples/mcp-demo/samyama-mcp-demo.mp4

F-13 Inventory / provenanceCWE-1104Layer L0

sdk/python/examples/mcp-demo/samyama-mcp-demo.mp4

Why: 1347369 bytes not read
Fix: —

INFOInventory / provenanceinv.oversize · src/query/executor/operator.rs

F-13 Inventory / provenanceCWE-1104Layer L0

src/query/executor/operator.rs

Why: 1136384 bytes not read
Fix: —

Tools it exposes

12 declared in source · 0 destructive

10 read2 write0 destructive

Tool	What it does	Risk
`bad_tool`	This tries to write	write
`cypher_query`	Execute a read-only Cypher query against the graph.	write
`count_nodes`	Count all nodes	read
`drug_neighbours`	Drugs that share a pathway (4-hop scan — intentionally slow)	read
`heavy-skew`	one value on ~50% of nodes, long tail — social/biomedical shape	read
`list_labels`	Aggregate nodes by label	read
`mild-skew`	10 values, most common 3x the rarest	read
`rare_probe`	Look up a specific rare drug by exact name	read
`sample_drugs`	List drug names	read
`schema_info`	Return the full graph schema including node types, edge types,	read
`top_people`	Top people by some metric	read
`uniform`	10 values, equal counts — the distribution the estimator assumes	read

Safer alternatives

All memory & knowledge servers →

Memento✓SAFE

gannonh/memento-mcp

Memento MCP: A Knowledge Graph Memory System for LLMs

4 destructiveGrade B

HiveMem✓SAFE

visterion/HiveMem

Personal knowledge system — MCP server with PostgreSQL semantic search and temporal knowledge graph

1 readGrade B

Ourmem✓SAFE

ourmem/omem

Shared Memory That Never Forgets — persistent memory for AI agents with Space-based sharing across agents and teams. Plugins for OpenCode, Claude Code, OpenClaw, MCP Server.

3 writeGrade B

Audit history

2026-10-06
✕BLOCK D 63 / 100
Score 67 -> 63 · dca807886561
2026-09-17
✕BLOCK D 67 / 100
First audit · bba296964a8f

README badge

Badges show the current verdict, BLOCK included

A badge cannot be hidden by the maintainer while the verdict is BLOCK. Fix the findings, push, and the re-audit updates it.

Questions about samyama-graph MCP server

What is the samyama-graph MCP server?

Rust graph-vector database: OpenCypher (99.9% of evaluated TCK scenarios pass), vector search, graph algorithms, RESP + HTTP. LDBC SNB Interactive and BI complete; 1B edges on one machine.

What tools does samyama-graph expose?

12 in total: 10 read-only, 2 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.

Is samyama-graph safe to connect to an agent?

No — not without reading the findings first. The audit graded it D (63/100) and found 3 critical or high issues in the source. Each one is listed on this page with the file and line it is on.

What credentials does samyama-graph need?

No credential environment variables were found in its source, so it appears to need none.

How current is this page?

The grade is for one exact copy of the source (dca807886561), read on 2026-10-06. The repository is watched and re-audited when it changes.

I already installed samyama-graph. What should I do?

Remove it from your client's MCP configuration, then rotate any API keys and tokens that were in your environment while it ran.

Can the verdict change?

Yes. The repository is watched and re-audited when its source changes. If the findings are removed the verdict updates, and the history on this page shows it.

Sources

samyama-ai/samyama-graph on GitHub — the source we read, its licence, stars and activityObserved · 6 Oct 2026
ModelsAtlas trust audit — BLOCK verdict for commit dca807886561Observed · 6 Oct 2026

Spotted something wrong? Report a correction · Methodology

Keep going in your AI

Opens with this page as the source. We add nothing else to the prompt.

What is the samyama-graph MCP server?
GChatGPT CClaude PPerplexity
What tools does samyama-graph expose?
GChatGPT CClaude PPerplexity
Is samyama-graph safe to connect to an agent?
GChatGPT CClaude PPerplexity