asset-aware-mcpCAUTION
Agent document collaboration: versioned PDF pages/annotations/forms, raster frames, DOCX/PPTX, spreadsheets and CSV/TSV, with checked edits, source provenance, CSL citations and wikilink evidence libraries
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
Citation-ready document infrastructure for AI agents: editable native documents, reusable evidence, and portable Foam/LightRAG wikis.
[](https://opensource.org/licenses/Apache-2.0)
🌐 繁體中文 · Docs Site · GitHub Wiki
v1.4.1 Consolidated native-document release
This release brings together the document and evidence work below. We keep the 1.4.x line and consolidate verified changes instead of releasing each feature. Check GitHub Releases for publication status and the changelog for details.
Discover the installed runtime's complete contract and enabled flags before acting. Since v1.4.0, native-contract-v2 can page schemas: inspect schema_delivery, select for_op, or follow the hash-pinned schema_request. Follow contract_request for paged policies too; abbrev
a0ad6eb4e1d4OBSERVED · 2026-10-12Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control.
claude mcp add asset-aware-mcp -- npx -y [email protected]
{
"mcpServers": {
"asset-aware-mcp": {
"command": "npx",
"args": [
"-y",
"[email protected]"
]
}
}
}Exposed tools (61)
52 read · 7 write · 2 destructive. Blast radius: 2 tools can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.
| Tool | Risk | Description |
|---|---|---|
cancel_job | read | |
citation_bundle | read | |
consult_knowledge_graph | read | |
convert_document | read | |
convert_docx_to_doc | read | |
convert_docx_to_odt | read | |
convert_docx_to_pdf | read | |
convert_pdf_to_docx | read | |
convert_pdf_to_pptx | read | |
delete_document | destructive | |
delete_docx | destructive | |
detect_etl_profile | read | |
discover_sources | read | |
document_asset | read | Consolidated document asset and section entrypoint. |
docx | read | |
docx_chart_data | read | |
docx_table | read | |
docx_table_edit_plan | write | |
docx_table_from_context | read | |
docx_table_to_context | read | |
docx_validate_roundtrip | read | |
etl_profile | read | |
export_document_segmentation | write | 匯出整合 manifest、blocks、assets、reading order 的 segmentation schema。 |
export_knowledge_graph | read | |
export_markdown | read | |
fetch_document_asset | read | |
find_evidence_spans | read | |
get_current_etl_profile | read | |
get_docx_content | read | |
get_etl_profile | read | |
get_job_status | read | |
get_section_blocks | read | |
get_section_content | read | |
get_section_detail | read | |
ingest_documents | read | |
ingest_docx | read | |
inspect_document_manifest | read | |
job | read | |
knowledge | read | |
list_documents | read | |
list_docx_blocks | read | |
list_docx_documents | read | |
list_etl_profiles | read | |
list_jobs | read | |
list_section_tree | read | |
load_etl_profile_from_json | read | |
ocr_pdf_document | write | Create a background OCR ingest job instead of blocking the MCP request. |
parse_pdf_structure | write | Create a background job using the configured structured PDF extractor. |
plan_table | read | |
save_docx | write | |
search_sections | read | |
search_source_location | read | |
section | read | Consolidated section navigation entrypoint. |
set_etl_profile | write | |
table_cite | read | |
table_data | read | |
table_draft | read | |
table_history | read | |
table_manage | read | |
verify_citation_ref | read | Verify a span-level AssetRef against the persisted citation index. |
visualize_document_layout | write | 產生 PDF page overlay,直接檢查 block bbox、類型與 reading order。 |
Trust audit
CAUTIONgrade C · trust 74/100 Install with care. The audit found things worth knowing before you trust its output.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | WARN |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | WARN |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (6 observation(s))
- Network
- declared (8 observation(s))
- Shell
- declared (1 observation(s))
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (25)
__import__(module_name, fromlist=["*"])
this.sys.print(INDENT + "---> Token: " + token.name + " (" + token.tokentype + ") in " + this.tmp.area + ", " + this.output.current.mystack.length);delete_document, delete_docx
.clineignore
.pre-commit-config.yaml
.nojekyll
.vscodeignore
randomizer = random.Random(243) # noqa: S311 -- reproducible non-secret fixture
("missing.png" if fault == "missing" else "../../../../outside.png"),const extensionDevelopmentPath = path.resolve(__dirname, '../../');
import { installAssistantAssets } from '../../assistantAssets';import { installClineMcpServer, isClineInstalled, removeClineMcpServer } from '../../clineMcpConfig';} from '../../codexMcpConfig';
http://127.0.0.1:8876/#/native-file-assets and release-testing, desktop 1440x1000 and
assert.strictEqual(__test__.isSafeOpenRouterUrl('http://127.0.0.2:8080/v1'), true);binary = base64.b64decode(
raw = base64.b64decode(blocks[0]["data"], validate=True)
base64.b64decode(images[0]["data"], validate=True),
png = base64.b64decode(blocks[0]["data"], validate=True)
raw = base64.b64decode(blocks[0]["data"], validate=True)
assert metadata["char_range"] == [0, len("Alpha βeta")]assert metadata["byte_range"] == [0, len("Alpha βeta".encode())]for bad in ([{"Count": "003", "Unit": "μg/mL"}], [{"Count": "3", "Unit": "μg/mL"}]):(2, "Unit", "μg/mL"),
("μg", "μg"),Gates applied: no_behavioural_pass.
a0ad6eb4e1d4full audit observations/trust-audit/mcp-server/u9401066__asset-aware-mcp.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-12 | a0ad6eb4e1d4 | CAUTION | C | 74 | first audit |
Questions
What is the asset-aware-mcp MCP server?
Agent document collaboration: versioned PDF pages/annotations/forms, raster frames, DOCX/PPTX, spreadsheets and CSV/TSV, with checked edits, source provenance, CSL citations and wikilink evidence libraries
What tools does asset-aware-mcp expose?
61 in total: 52 read-only, 7 that write, and 2 that can delete or overwrite (delete_document, delete_docx). Every one is listed on this page with its risk.
Is asset-aware-mcp safe to connect to an agent?
With care. The audit graded it C (74/100) and found 25 things worth knowing before you trust this server, listed below with the exact line each was found on. Separately from the audit: 2 of its tools can destroy data, so scope the token you give it to what you actually need.
What credentials does asset-aware-mcp need?
No credential environment variables were found in its source, so it appears to need none.
How does asset-aware-mcp run?
It speaks stdio, so it runs as a local process your client starts. It is published on npm as asset-aware-mcp at 1.4.1.
How current is this page?
The grade is for one exact copy of the source (a0ad6eb4e1d4), read on 2026-10-12. The repository is watched and re-audited when it changes.