Live survey304 models · 14 providers · 23,592 auditsAudited implement · SAFE →What changedHow we verifyllms.txt

MCP server · Audited from source

mibayy-token-savior MCP server

by mibayy/token-saviorMIT licenceAudited 12 Nov 2026 · source 66942c22e407

PyPI token-savior-recall 4.19.0stdio★ 1,217
!CAUTIONGrade F · 58/100 · 12 Nov 2026 · 66942c2
1,217GitHub stars

Is it safe? With care — CAUTION, grade F (58/100). No tools were found in its source. The audit recorded 27 findings (0 critical, 0 high, 11 medium, 15 low). Maintainer's description: MCP server that gets Claude to 97.9% (188/192) on a real coding benchmark at -80% active tokens and -83% wall time, vs 78.3% plain. Structural code navigation + persistent memory engine. Works with every MCP client.

Ask an AI about this pageGChatGPTCClaudePPerplexityGeGeminiXGrok
More
Cite

Cite this page

ModelsAtlas. “mibayy-token-savior MCP server (mibayy/token-savior)”. Retrieved 12 Nov 2026. https://modelsatlas.com/mcp/mibayy/token-savior

Trust audit

Audit 0.4.1 · full source read 12 Nov 2026

Trust score58/100!CAUTION F
0 critical0 high11 medium15 low1 info

27 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

  • FilesystemDeclared9 observations
  • NetworkDeclared9 observations
  • ShellDeclared2 observations
  • DependenciesPinnedNothing floats
  • SecretsFoundSee the findings

“Declared” means the server says it does this and the source agrees. Browser, file and shell access are expected of some kinds of server.

Five audit layers

LayerWhat we checkResult
L0 ProvenanceRepo owner, licence, archive status, activityPASS
L1 Static sourceSecrets, path traversal, supply chainWARN 24 findings
L2 InstructionsPrompt injection and authority framing in docs and tool descriptionsPASS 1 finding
L3 InventoryDeclared tools match the sourcePASS
L4 SandboxRun in isolation and watch behaviourSKIPPED

Provenance & runtime

Who published it, and what the audit could and could not run

Provenance

  • Repository owner mibayyUser accountAccount type recorded when we read the repository.
  • PackagePublished on PyPItoken-savior-recall at version 4.19.0
  • LicenceMITFrom the repository, as read at the audited commit.
  • Last push28 days agoActivity of the repository when we read it.
  • Source commit read66942c22e407The audit covers exactly this commit, read on 12 Nov 2026.

Runtime

  • Sandbox runNot run yetThis audit read the source only; it did not execute the server.
  • Tool inventoryNone declaredNo tools were found in the source.

Every value here is read from the record's own repository and audit.Suggest a correction

Findings (27)

Sorted by severity · the 25 most severe of 27 are listed

MEDIUMCode injectioncode.dynamic_import · src/token_savior/library_api.py:315
F-02 Code injectionCWE-78 · CWE-94 · CWE-95Layer L1
mod = importlib.import_module(mod_name)
Why
No further detail recorded for this rule.
Fix
—
MEDIUMCode injectioncode.dynamic_import · src/token_savior/library_api.py:366
F-02 Code injectionCWE-78 · CWE-94 · CWE-95Layer L1
mod = importlib.import_module(module)
Why
No further detail recorded for this rule.
Fix
—
MEDIUMInformation disclosuredisclose.log_secret · hooks/memory-session-stop.sh:448
F-12 Information disclosureCWE-209 · CWE-532Layer L1
print(f\"  · '{s['token']}' ×{s['count']} — consider memory_save\", file=sys.stderr)
Why
No further detail recorded for this rule.
Fix
—
MEDIUMNetwork egressnet.raw_ip · src/token_savior/memory/viewer.py:104
F-04 Network egressCWE-200 · CWE-319Layer L1
_logger.info("[token-savior:viewer] listening on http://127.0.0.1:%s", port)
Why
No further detail recorded for this rule.
Fix
—
MEDIUMNetwork egressnet.raw_ip · src/token_savior/memory/viewer.py:158
F-04 Network egressCWE-200 · CWE-319Layer L1
f"http://127.0.0.1:{port}/status", timeout=1.0,
Why
No further detail recorded for this rule.
Fix
—
MEDIUMNetwork egressnet.raw_ip · src/token_savior/server_handlers/memory.py:461
F-04 Network egressCWE-200 · CWE-319Layer L1
lines.append(f"🌐 Viewer: ok — http://127.0.0.1:{port}")
Why
No further detail recorded for this rule.
Fix
—
MEDIUMHard-coded secretssecret.generic · tests/test_config_analyzer.py:354
F-03 Hard-coded secretsCWE-798 · CWE-321Layer L1

Excerpt withheld: it is a secret, and we never print one.

Why
No further detail recorded for this rule.
Fix
—
MEDIUMHard-coded secretssecret.github · tests/test_config_analyzer.py:252
F-03 Hard-coded secretsCWE-798 · CWE-321Layer L1

Excerpt withheld: it is a secret, and we never print one.

Why
No further detail recorded for this rule.
Fix
—
MEDIUMHard-coded secretssecret.github · tests/test_config_analyzer.py:809
F-03 Hard-coded secretsCWE-798 · CWE-321Layer L1

Excerpt withheld: it is a secret, and we never print one.

Why
No further detail recorded for this rule.
Fix
—
MEDIUMHard-coded secretssecret.github · tests/test_config_analyzer.py:824
F-03 Hard-coded secretsCWE-798 · CWE-321Layer L1

Excerpt withheld: it is a secret, and we never print one.

Why
No further detail recorded for this rule.
Fix
—
MEDIUMHard-coded secretssecret.private_key · tests/test_config_analyzer.py:261
F-03 Hard-coded secretsCWE-798 · CWE-321Layer L1

Excerpt withheld: it is a secret, and we never print one.

Why
No further detail recorded for this rule.
Fix
—
LOWCode injectioncode.dynamic_import · tests/test_memory_vector_setup.py:122
F-02 Code injectionCWE-78 · CWE-94 · CWE-95Layer L1
return importlib.__import__(name, *args, **kwargs)
Why
No further detail recorded for this rule.
Fix
—
LOWInsecure cryptocrypto.weak_hash · hooks/memory-session-start.sh:190
F-09 Insecure cryptoCWE-327 · CWE-338Layer L1
current_hash = hashlib.md5(payload.encode()).hexdigest()
Why
No further detail recorded for this rule.
Fix
—
LOWInsecure cryptocrypto.weak_hash · src/token_savior/checkpoint_ops.py:17
F-09 Insecure cryptoCWE-327 · CWE-338Layer L1
checkpoint_id = hashlib.md5(seed.encode()).hexdigest()[:12]
Why
No further detail recorded for this rule.
Fix
—
LOWInsecure cryptocrypto.weak_hash · src/token_savior/memory/symbol_embeddings.py:101
F-09 Insecure cryptoCWE-327 · CWE-338Layer L1
content_hash = hashlib.sha1(
Why
No further detail recorded for this rule.
Fix
—
LOWInsecure cryptocrypto.weak_hash · src/token_savior/memory/symbol_embeddings.py:241
F-09 Insecure cryptoCWE-327 · CWE-338Layer L1
"content_hash": hashlib.sha1(
Why
No further detail recorded for this rule.
Fix
—
LOWInsecure cryptocrypto.weak_hash · src/token_savior/memory/tool_capture.py:64
F-09 Insecure cryptoCWE-327 · CWE-338Layer L1
return hashlib.sha1(s.encode("utf-8", errors="replace")).hexdigest()[:16]
Why
No further detail recorded for this rule.
Fix
—
LOWFilesystem / pathfs.traversal · tests/maison/test_05_integrite.py:177
F-05 Filesystem / pathCWE-22 · CWE-59Layer L1
["../evasion.py", "/etc/passwd", "pkg/../../evasion.py"],
Why
No further detail recorded for this rule.
Fix
—
LOWFilesystem / pathfs.traversal · tests/test_arg_aliases_and_resolution.py:139
F-05 Filesystem / pathCWE-22 · CWE-59Layer L1
(wt / ".git").write_text("gitdir: ../../../.git/worktrees/fix-98\n")
Why
No further detail recorded for this rule.
Fix
—
LOWFilesystem / pathfs.traversal · tests/test_autodiscover_roots.py:215
F-05 Filesystem / pathCWE-22 · CWE-59Layer L1
(wt / ".git").write_text("gitdir: ../../../.git/worktrees/fix-98\n")
Why
No further detail recorded for this rule.
Fix
—
LOWFilesystem / pathfs.traversal · tests/test_docker_analyzer.py:140
F-05 Filesystem / pathCWE-22 · CWE-59Layer L1
assert "../../" not in result
Why
No further detail recorded for this rule.
Fix
—
LOWFilesystem / pathfs.traversal · tests/test_server_integration.py:200
F-05 Filesystem / pathCWE-22 · CWE-59Layer L1
{"file_paths": ["../../../etc/passwd"], "label": "evil"},
Why
No further detail recorded for this rule.
Fix
—
LOWNetwork egressnet.raw_ip · tests/test_memory_viewer.py:44
F-04 Network egressCWE-200 · CWE-319Layer L1
f"http://127.0.0.1:{port}/status", timeout=0.5,
Why
No further detail recorded for this rule.
Fix
—
LOWNetwork egressnet.raw_ip · tests/test_memory_viewer.py:122
F-04 Network egressCWE-200 · CWE-319Layer L1
with urllib.request.urlopen(f"http://127.0.0.1:{port}/") as r:
Why
No further detail recorded for this rule.
Fix
—
LOWPrompt injectionprompt.credential_read · CLAUDE.md:41
F-01 Prompt injectionCWE-94 · CWE-1427Layer L2
`Edit` / `Write` / `Read` / `Grep` stay allowed for `.env`, `.yml`, `.json`
Why
asks the agent to read credentials
Fix
—

Tools it exposes

None declared in source

No tools were found in the source, so we cannot say what this server exposes.

Install

Package version 4.19.0 · audited commit 66942c2

Read the findings first. The most severe is Code injection (medium) at src/token_savior/library_api.py.

No connect command is recorded for this server yet. Follow the repository's own instructions, and read the findings above first.

Audit history

  1. !CAUTION F 58 / 100

    First audit · 66942c22e407

README badge

For maintainers · updates on every re-audit

ModelsAtlas: CAUTION FMODELSATLASCAUTION · F
[![ModelsAtlas: CAUTION F](https://modelsatlas.com/mcp/mibayy/token-savior/badge.svg)](https://modelsatlas.com/mcp/mibayy/token-savior/)

The badge always shows the latest audit; a maintainer cannot pin an old verdict.

Other ratings

Cited, never blended into our score

No third-party security ratings are recorded for this server yet. When they exist they appear here with their source and date, beside, not inside, our grade.

Questions about mibayy-token-savior MCP server

What is the mibayy-token-savior MCP server?
MCP server that gets Claude to 97.9% (188/192) on a real coding benchmark at -80% active tokens and -83% wall time, vs 78.3% plain. Structural code navigation + persistent memory engine. Works with every MCP client.
Is mibayy-token-savior safe to connect to an agent?
With care. The audit graded it F (58/100) and found 25 things worth knowing before you trust this server, listed below with the exact line each was found on.
What credentials does mibayy-token-savior need?
It reads API_KEY, KEY, MY_KEY, STRIPE_KEY, TELEGRAM_BOT_TOKEN, TOKEN_SAVIOR_AUTODISCOVER, TOKEN_SAVIOR_BANNER, TOKEN_SAVIOR_CHAIN_NUDGE, TOKEN_SAVIOR_CLIENT, TOKEN_SAVIOR_DASHBOARD_HOST, TOKEN_SAVIOR_DASHBOARD_PORT and TOKEN_SAVIOR_DATA_DIR from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does mibayy-token-savior run?
It speaks stdio, so it runs as a local process your client starts. It is published on PyPI as token-savior-recall.
How current is this page?
The grade is for one exact copy of the source (66942c22e407), read on 2026-11-12. The repository is watched and re-audited when it changes.

Sources

  • mibayy/token-savior on GitHub — the source we read, its licence, stars and activityObserved · 12 Nov 2026
  • ModelsAtlas trust audit — CAUTION verdict for commit 66942c22e407Observed · 12 Nov 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.