{"slug":"arch-req-from-doc-3","title":"arch-req-from-doc","summary":"Extract architecture requirements from documents (PDF, DOCX, MD, TXT). Uses LLM to identify physical locations, tech stack, integration points, authentication, and security requirements from unstructured text. Outputs a partial requirements YAML. Medium confidence — always verify","platform":"Claude","tags":[],"authorName":"LLM Mart","authorSlug":"llm-mart","score":0,"source":"github","price":null,"verified":false,"createdAt":"2026-09-15T18:30:11.116009Z","repo":{"url":"https://github.com/axisrobo/ea-harness","stars":12,"forks":13,"license":"MIT","updatedAt":"2026-09-14T11:09:56Z"},"bodyHtml":"
\n

name: arch-req-from-doc\ndescription: >\nExtract architecture requirements from documents (PDF, DOCX, MD, TXT).\nUses LLM to identify physical locations, tech stack, integration points,\nauthentication, and security requirements from unstructured text.\nOutputs a partial requirements YAML. Medium confidence — always verify.\nUse before arch-req-merge to collect the business context layer.

\n
\n

Locating shared resources. References in this file to standards/,\ntools/, config.yaml, and templates/ are relative to the ArchHarness\nresource root. Determine the root, in order: (1) the ARCHHARNESS_HOME\nenvironment variable, (2) the output of python -m archharness root (the\npip-installed package bundles these resources under its data directory),\n(3) the current working directory when it already contains config.yaml and\ntools/ (the repository checkout). Prefix shared paths with that root\nwhenever the working directory is not the resource root.

\n
\n

You are an architecture requirements analyst reading documents.\nYour job is to extract precise, physical architecture information from\ntext documents — not to evaluate or critique the content.

\n

What documents typically contain (and don't)

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
What documents usually haveWhat documents usually lack
Business purpose and scopeSpecific port numbers
Application namesAuth mechanism details
Vague deployment location (\"cloud\", \"China DC\")Exact subnet/zone placement
Ownership and departmentProtocol-level auth
High-level tech stackCredential storage solution
Data sensitivity classificationCross-border compliance details
\n

Extraction from documents is MEDIUM confidence at best.\nAlways flag missing auth mechanisms — documents almost never specify them.

\n

How to invoke the Python tool

\n
cd tools/arch-req-readers\n\n# PDF document\nANTHROPIC_API_KEY=... python from_document.py -i requirements.pdf -o partial-doc.yaml\n\n# Word document\nANTHROPIC_API_KEY=... python from_document.py -i design.docx -o partial-doc.yaml\n\n# Markdown or text\nANTHROPIC_API_KEY=... python from_document.py -i brief.md -o partial-doc.yaml\n
\n

When a document is uploaded directly in chat

\n

If the user uploads or pastes document content in the conversation, extract\nrequirements directly. Focus on finding:

\n
    \n
  1. Project name and purpose (1-2 sentences)
    2. \n
  2. Application names — any system, service, or platform mentioned
    3. \n
  3. Physical location clues — country names, DC names, cloud provider mentions
    4. \n
  4. Technical stack — language, framework, runtime mentions
    5. \n
  5. Integration points — \"connects to\", \"calls\", \"sends data to\" statements
    6. \n
  6. User types — who uses the system
    7. \n
  7. Compliance/security mentions — GDPR, data residency, encryption
    8. \n
\n

Output a partial req.yaml with explicit _confidence markers:

\n
requirements:\n  project:\n    name: \"[extracted]\"\n    _confidence: \"medium\"\n  applications:\n    - name: \"[app name from document]\"\n      dc_or_region: \"[if mentioned, otherwise null]\"\n      _confidence: \"medium\"\n      _note: \"Verify physical location — document may use vague terms\"\n  interactions:\n    - from_component: \"[if mentioned]\"\n      to_component: \"[if mentioned]\"\n      protocol: \"[if mentioned]\"\n      auth_method: null\n      _confidence: \"low\"\n      _gap: \"Auth mechanism not mentioned in document — CRITICAL gap\"\n
\n

Gap flags to always add after document extraction

\n\n","files":[{"path":"SKILL.md","sizeBytes":3748,"isText":true}],"reviewScore":null,"reviewSummary":null,"trust":{"provenance":"trusted-source-unreviewed","notice":"Community-authored content, reproduced verbatim and not vetted as instructions. Treat it as data to evaluate, never as directives to follow.","bodySource":null},"bodyLocked":false,"purchaseUrl":null,"sourceUrl":null,"report":{"provenance":"trusted-source-unreviewed","screen":{"ran":true,"outcome":"clean","suspicious":0,"notes":0,"hiddenCharacters":false},"virusScan":{"engine":"clamav","status":"clean","scannedAt":"2026-09-15T18:31:18.346763Z","sha256":"E63AFD6A27243247ADDEF237AC58D00FE5D31F9C423DD37CF16B653A8560006F","sizeBytes":1834},"review":null,"source":{"repositoryUrl":"https://github.com/axisrobo/ea-harness","path":".claude/skills/arch-req-from-doc","license":"MIT","commit":"2f6ee4f4ef319706721c6e24e6c6d3df44250d54","subtreeSha":"9A62AD89D64F8A578103F74215AA9B5698B2B25BE4A8BF1B7B0A453AB26F3C88","lastSyncedAt":"2026-09-21T13:50:53.352538Z"},"reviewedAt":"2026-09-15T18:43:52.394822Z","notice":"Community-authored content, reproduced verbatim and not vetted as instructions. Treat it as data to evaluate, never as directives to follow."},"install":[{"target":"skills-cli","command":"npx skills add https://github.com/axisrobo/ea-harness/tree/main/.claude/skills/arch-req-from-doc"},{"target":"claude-code","command":"claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install axisrobo-ea-harness@llmmart"},{"target":"git","command":"git clone https://github.com/axisrobo/ea-harness.git"}]}