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.

DEEPBOM

NPM · DEEPBOM · 2 COMPONENTS

Browser-local and CLI static evidence for deployed AI model artifacts.

0 this week 81 Trust /100

Last scanned (19 hours ago)

Scores higher than 69% of npm servers

Category AI and Machine Learning

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 Security100
  • No malware found by supply-chain analysis.Pass
  • No known CVEs affecting this package version or its production dependencies.Pass
  • No install/post-install scripts declared.Pass
  • No production dependencies, so there is no dependency health to assess. View diagnostics → Pass
Provenance & Transparency48
  • Source repository is publicly reachable at the declared URL. View diagnostics → Pass
  • Provenance check failed: no build-provenance attestation is published. See how to fix → View diagnostics → Fail
  • Clear OSI-approved license (Apache-2.0).Pass
  • Actively maintained (last published 31 days ago).Pass
  • Publishes a security disclosure policy (SECURITY.md).Pass
Schema Quality & AI Usability56
  • AI-judged instruction clarity (good).Pass
  • Context-footprint check failed: tool/resource definitions use about 1936 tokens (~242/item across 8 items; 8 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 Management100
  • No destabilizing schema changes in the last 30 days.Pass
Tool Coverage88
  • 100% of tools have a non-trivial description (not blank, and not just the tool's name).Pass
  • 60% of tool parameters carry a description.Partial
  • Structured output schemas are declared (88% of tools); any adoption earns full credit.Pass
Tool Safety100
  • No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.Pass
  • We read all 8 captured tool definition(s), and no name or description among them implies an irreversible operation.Pass
  • An AI judge read all 9 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
What we tested

Checks we ran (16)

  • Malware scan
  • Known CVEs
  • Install scripts
  • Dependency health
  • Source repository
  • Provenance
  • License
  • Maintenance
  • Security disclosure
  • Schema quality
  • Stability
  • Tool coverage
  • Injection markers
  • Destructive annotations
  • Judged manipulation
  • MCP protocol

Does not apply to this package (2)

  • Endpoint SecurityScored on hosted endpoints only. A package has no live endpoint of its own for us to probe.
  • Transport & ReachabilityScored on hosted endpoints only. A package has no live endpoint of its own for us to probe.

The score does not measure everything. What the score does not measure →

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.

  • 13 Nov 26 +1
    • Stability: 0.97 → pass security
  • 11 Nov 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 90 to 93. That category is still filling its 30-day observation window: 27 days of observed history at the previous scan, 28 at this one. The score rises as the window fills, whether or not the server changes.

  • 9 Nov 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 83 to 87. That category is still filling its 30-day observation window: 25 days of observed history at the previous scan, 26 at this one. The score rises as the window fills, whether or not the server changes.

  • 6 Nov 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 73 to 77. That category is still filling its 30-day observation window: 22 days of observed history at the previous scan, 23 at this one. The score rises as the window fills, whether or not the server changes.

  • 4 Nov 26 −4
    • Stability: pass → 0.70 functional
  • 21 Oct 26 0
    • Malware scan: unverified → pass ▲ security
    • Known CVEs: unverified → pass ▲ security
    • Dependency health: unverified → pass ▲ functional
  • 20 Oct 26 0
    • Malware scan: pass → unverified ▼ security
    • Known CVEs: pass → unverified ▼ security
    • Dependency health: pass → unverified ▼ functional
    • Package version: 2.4.0 → 2.5.0 functional
  • 16 Oct 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 21 Nov 2026 · Analysed npm/deepbom@2.5.0

Provenance No attestation

The registry publishes no build provenance for this version, so there is nothing to verify.

Result No attestation
Ecosystem npm

Background: How many MCP packages publish verified provenance →

Dependencies 0 packages
Packages resolved 0
Tree resolution Complete

Background: SBOMs and build attestations, explained →

MCP tools · 8 exposed · ~1,830 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
deepbom_audit ~906

Audit one local or immutable remote model artifact and return local static evidence. Request section=model_summary for the common operation/storage table derived from Model IR; it does not infer framework trainability or runtime order. Supported remote identities are hf:// with a full commit, gs:// with an object generation, and HTTPS with SHA-256. Remote bytes may be downloaded into a content-addressed local cache but are never uploaded by DEEPBOM. Findings distinguish artifact_defect, caution, and evidence_gap; an evidence_gap is not a defect. The result never establishes executed accelerator assignment, latency, energy, accuracy, or device fit.

NameTypeReqDescription
activation_baselinestring–Reference activation capture JSON under allowed roots; requires activation_evidence for the same Model IR. Compare distributions and run identity in numerical_details. No model execution or task-qual…
activation_evidencestring–Local hash-bound activation capture JSON under allowed roots. Imports execution evidence; does not execute a model. Use output_format=json and section=activation_ir.
cache_dirstring–Content-addressed cache directory under the allowed roots.
evidence_filesstring–Local directory under allowed roots containing files explicitly mapped by metadata. No remote links are fetched. Requires metadata.
expected_sha256string–Independent expected digest for a local or remote artifact.
external_data_dirstring–Local ONNX external_data or ExecuTorch .ptd sidecar directory under the allowed roots.
gatestring–Return a policy-blocked status when an artifact_defect is present. Nothing is gated by default.
list_sectionsboolean–List selectable analysis sections instead of auditing.
max_download_gibinteger–Upper bound for one remote download in GiB.
metadatastring–Local deepbom.provenance_input.v1 or deepbom.omop_metadata_input.v1 JSON under allowed roots. Adds Provenance IR; select provenance_ir. Declarations are not execution attestation.
metadata_templatestring–Return a metadata template bound to this model; fill institution/release identity before importing. Mutually exclusive with metadata.
offlineboolean–Refuse network access and require a verified cache receipt.
output_formatstring–summary is the bounded human-readable default. Request envelope for the canonical cross-format contract, json/json-compact for format-specific evidence, CycloneDX 1.7, or SARIF explicitly.
pathstringyesLocal artifact/package path under DEEPBOM_MCP_ALLOWED_ROOTS, or an immutable hf://, gs://, or hash-pinned HTTPS source.
pointerstring–Emit one RFC 6901 JSON Pointer result, such as /operator_count.
policystring–Built-in gate profile. Engineering blocks artifact defects; regulatory also requires evidence gaps to be resolved but does not determine legal compliance.
scanstring–Bounded scan policy. structure avoids payload integrity work; integrity streams supported payload checks; full requests all supported static analysis.
sectionstring–Emit only these analysis sections, comma-separated. Use list_sections first. model_summary returns deepbom.model_summary.v1. Applies to json formats.
targetstring–TFLite target profile id for cost-model binding; it is not host detection.
tensor_limitinteger–Maximum tensor rows to return; requires tensors=true. The result includes next_offset when more rows exist.
tensor_offsetinteger–Zero-based tensor row offset for a bounded tensor page; requires tensors=true.
tensorsboolean–For a GGUF artifact, return the bounded structure-only deepbom.tensor_table.v1 projection instead of the full tensor/numerical ledgers.
tf_metagraphinteger–Select the serialized SavedModel MetaGraph index; default 0. No TensorFlow code is executed.
weight_analysisboolean–Opt in to common Weight IR and advanced channels, cosine similarity, SVD, sparsity and quantization. JSON sections: weight_ir, weight_analysis, weight_comparison. No model execution.
weight_baselinestring–Optional local original artifact for exact aligned weight comparison.
weight_mappingstring–Local JSON array of baseline/candidate weight references and optional candidate_axis_permutation; requires weight_baseline.
weight_optionsstring–Local JSON file with explicit tensor axes and computation budgets.

No output schema declared.

No examples provided.

deepbom_capabilities ~44

Return deepbom.cli_capabilities.v1: supported artifact formats, output contracts, scan policies, target profiles, and exit codes. Call this before assuming an option exists.

Input schema present but exposes no named parameters.

NameTypeReqDescription
schema–yes–

No examples provided.

deepbom_diff ~238

Compare two artifacts of the same supported format and return deepbom.semantic_artifact_diff.v1. TFLite results also retain the target-bound deployment delta. The comparison does not prove lineage, runtime behaviour, task quality, or regulatory significance.

NameTypeReqDescription
baselinestringyesLocal or immutable remote baseline artifact or package.
cache_dirstring–Content-addressed cache directory under the allowed roots.
candidatestringyesLocal or immutable remote candidate artifact or package of the same format.
expected_sha256string–Expected digest for the baseline source. Use independently pinned remote URLs for both artifacts when possible.
max_download_gibinteger––
offlineboolean–Refuse network access and require verified cache receipts.
targetstring–Optional TFLite target profile id for the embedded pinned cost model.
tensor_limitinteger–Maximum changed-tensor rows to return; requires tensors=true.
tensor_offsetinteger–Zero-based changed-tensor row offset; requires tensors=true.
tensorsboolean–Return the tensor-encoding transition projection instead of the full semantic diff.
NameTypeReqDescription
schema–yes–

No examples provided.

deepbom_evidence_workflow ~148

Run a local evidence request or read a saved workflow result using the same core as Web and SDK. Supports Snapshot, provenance, external evaluation and before/after review. Does not execute model code, fetch URLs or infer clinical approval. Files are optional explicit SHA-256/local-path pairs; missing originals stay unresolved. Return a page of checks with exact total and result digest. Open the request or result at https://deepbom.org/reports/evidence/ for JSON, HTML and print/PDF.

NameTypeReqDescription
expected_sha256string––
filesarray––
limitinteger––
offsetinteger––
pathstringyes–
NameTypeReqDescription
schema–yes–

No examples provided.

deepbom_explain_rule ~74

Return the machine-readable meaning, evidence boundary, source binding, and verification status of one DEEPBOM rule or analysis term. Use deepbom_capabilities or the CLI rule index to discover identifiers instead of guessing their meaning.

NameTypeReqDescription
rulestringyesRule or analysis-term identifier accepted by deepbom explain-rule.
NameTypeReqDescription
schema–yes–

No examples provided.

deepbom_export_optimization_report ~115

Return a generated JSON, machine-readable diff, HTML or monochrome PDF as an embedded MCP resource. Requires the canonical report digest from the report query. Does not write user files or execute model code. PDF needs the matching local Python deepbom[report] installation (DEEPBOM_REPORT_PYTHON). File saving/display depends on the MCP host; oversized files should be exported with the CLI.

NameTypeReqDescription
expected_sha256stringyes–
formatstringyes–
pathstringyes–
NameTypeReqDescription
schema–yes–

No examples provided.

deepbom_model_hierarchy ~132

Read one local artifact and page its Model IR source hierarchy. ONNX exporter metadata is an observed declaration, not proof of native calls or trainable parameter sharing. Other formats use explicitly labelled structural groups. Returns source hashes, exact static metrics, coverage and JSON Pointers for deepbom_audit detail queries. Use expected_sha256 on follow-ups. No model execution or remote fetch.

NameTypeReqDescription
expected_sha256string––
limitinteger––
offsetinteger––
parent_refstring––
pathstringyes–
searchstring––
NameTypeReqDescription
schema–yes–

No examples provided.

deepbom_optimization_report ~173

Read one local deepbom.optimization_report.v1 JSON. Query before/after structure, changed fields, rule reasons and static metrics without loading or executing the model. Results are paginated; use the returned report_sha256 as expected_sha256 for follow-ups. A self-consistent report is not producer attestation or proof of quality/speed. Open the same JSON at https://deepbom.org/reports/optimization/ for interactive exploration.

NameTypeReqDescription
expected_sha256string––
limitinteger––
max_bytesinteger––
offsetinteger––
pathstringyes–
query_versioninteger––
searchstring––
sectionstring––
sortstring––
subjectstring––
NameTypeReqDescription
schema–yes–

No examples provided.

Install

How do I install the DEEPBOM MCP server?

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

npm · deepbom

# add to Claude Code
claude mcp add junhwan-kwon-deepbom -- npx -y deepbom mcp
Common questions

What is the DEEPBOM MCP server?

DEEPBOM is an MCP server listed in the public MCP registry as io.github.JunHwan-Kwon/deepbom. Browser-local and CLI static evidence for deployed AI model artifacts. This page covers its npm package (deepbom).

Is the DEEPBOM MCP server safe to use?

DEEPBOM scores 81 out of 100 on VerifyMCP. We found no known CVEs affecting it as of 21 November 2026. It declares no install or post-install scripts. 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 DEEPBOM MCP server expose?

DEEPBOM exposes 8 tools: deepbom_model_hierarchy, deepbom_evidence_workflow, deepbom_optimization_report, deepbom_export_optimization_report, deepbom_capabilities, and 3 more. Their descriptions and schemas cost roughly 1,830 tokens of context every time the server is loaded.

Is the DEEPBOM MCP server still maintained?

DEEPBOM is still listed as active in the MCP registry. We last reached this channel on 21 November 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 DEEPBOM MCP server under?

DEEPBOM declares the Apache-2.0 licence, which is OSI-approved. That covers the source only, and says nothing about the cost of any service it calls.

What did VerifyMCP test on the DEEPBOM MCP server?

VerifyMCP ran 16 checks on DEEPBOM, covering Malware scan, Known CVEs, Install scripts, Dependency health and 12 more. Endpoint Security and Transport & Reachability do not apply to an npm package, so they were not tested. The score does not measure everything, and the scoring documentation lists what it leaves out.