Decompose
Decompose any text or URL into classified semantic units. Each unit gets authority level, risk category, attention score, entity extraction, and irreducibility flags. No LLM required. Deterministic. Runs locally.
Setup
1. Install
pip install decompose-mcp
2. Configure MCP Server
Add to your OpenClaw MCP config:
{
"mcpServers": {
"decompose": {
"command": "python3",
"args": ["-m", "decompose", "--serve"]
}
}
}
3. Verify
python3 -m decompose --text "The contractor shall provide all materials per ASTM C150-20."
Available Tools
decompose_text
Decompose any text into classified semantic units.
Parameters:
text (required) — The text to decomposecompact (optional, default: false) — Omit zero-value fields for smaller outputchunk_size (optional, default: 2000) — Max characters per unit
Example prompt: "Decompose this spec and tell me which sections are mandatory"
Returns: JSON with units array. Each unit contains:
authority — mandatory, prohibitive, directive, permissive, conditional, informationalrisk — safety_critical, security, compliance, financial, contractual, advisory, informationalattention — 0.0 to 10.0 priority scoreactionable — whether someone needs to act on thisirreducible — whether content must be preserved verbatimentities — referenced standards and codes (ASTM, ASCE, IBC, OSHA, etc.)dates — extracted date referencesfinancial — extracted dollar amounts and percentagesheading_path — document structure hierarchy
decompose_url
Fetch a URL and decompose its content. Handles HTML, Markdown, and plain text.
Parameters:
url (required) — URL to fetch and decomposecompact (optional, default: false) — Omit zero-value fields
Example prompt: "Decompose https://spec.example.com/transport and show me the security requirements"
What It Detects
- Authority levels — RFC 2119 keywords: "shall" = mandatory, "should" = directive, "may" = permissive
- Risk categories — safety-critical, security, compliance, financial, contractual
- Attention scoring — authority weight x risk multiplier, 0-10 scale
- Standards references — ASTM, ASCE, IBC, OSHA, ACI, AISC, AWS, ISO, EN
- Financial values — dollar amounts, percentages, retainage, liquidated damages
- Dates — deadlines, milestones, notice periods
- Irreducibility — legal mandates, threshold values, formulas that cannot be paraphrased
Use Cases
- Pre-process documents before sending to your LLM — save 60-80% of context window
- Classify specs, contracts, policies, regulations by obligation level
- Extract standards references and compliance requirements
- Route high-attention content to specialized analysis chains
- Build structured training data from raw documents
Performance
- ~14ms average per document on Apple Silicon
- 1,000+ chars/ms throughput
- Zero API calls, zero cost, works offline
- Deterministic — same input always produces same output
Security & Trust
Text classification is fully local. The decompose_text tool performs all processing in-process with no network I/O. No data leaves your machine.
URL fetching performs outbound HTTP requests. The decompose_url tool fetches the target URL, which necessarily involves network I/O to the specified host. This is why the skill declares the network permission in claw.json. If you do not need URL fetching, you can use decompose_text exclusively with no network access required.
SSRF protection. URL fetching blocks private/internal IP ranges before connecting: 0.0.0.0/8, 10.0.0.0/8, 100.64.0.0/10, 127.0.0.0/8, 169.254.0.0/16, 172.16.0.0/12, 192.168.0.0/16, ::1/128, fc00::/7, fe80::/10. The implementation resolves the hostname via DNS before connecting and checks all returned addresses against the blocklist. See src/decompose/mcp_server.py lines 19-49.
No API keys or credentials required. No external services are contacted except when using decompose_url to fetch user-specified URLs.
Source code is fully auditable. The complete source is published at github.com/echology-io/decompose. The PyPI package is built from this repo via GitHub Actions (publish.yml) using PyPI Trusted Publishers (OIDC), so the published artifact is traceable to a specific commit.
Resources
1---2name: decompose-mcp3description: Decompose any text into classified semantic units — authority, risk, attention, entities. No LLM. Deterministic.4---5
6# Decompose
7
8Decompose any text or URL into classified semantic units. Each unit gets authority level, risk category, attention score, entity extraction, and irreducibility flags. No LLM required. Deterministic. Runs locally.
9
10## Setup
11
12### 1. Install
13
14```bash
15pip install decompose-mcp
16```
17
18### 2. Configure MCP Server
19
20Add to your OpenClaw MCP config:
21
22```json
23{
24 "mcpServers": {
25 "decompose": {
26 "command": "python3",
27 "args": ["-m", "decompose", "--serve"]
28 }
29 }
30}
31```
32
33### 3. Verify
34
35```bash
36python3 -m decompose --text "The contractor shall provide all materials per ASTM C150-20."
37```
38
39## Available Tools
40
41### `decompose_text`
42
43Decompose any text into classified semantic units.
44
45**Parameters:**
46- `text` (required) — The text to decompose
47- `compact` (optional, default: false) — Omit zero-value fields for smaller output
48- `chunk_size` (optional, default: 2000) — Max characters per unit
49
50**Example prompt:** "Decompose this spec and tell me which sections are mandatory"
51
52**Returns:** JSON with `units` array. Each unit contains:
53- `authority` — mandatory, prohibitive, directive, permissive, conditional, informational
54- `risk` — safety_critical, security, compliance, financial, contractual, advisory, informational
55- `attention` — 0.0 to 10.0 priority score
56- `actionable` — whether someone needs to act on this
57- `irreducible` — whether content must be preserved verbatim
58- `entities` — referenced standards and codes (ASTM, ASCE, IBC, OSHA, etc.)
59- `dates` — extracted date references
60- `financial` — extracted dollar amounts and percentages
61- `heading_path` — document structure hierarchy
62
63### `decompose_url`
64
65Fetch a URL and decompose its content. Handles HTML, Markdown, and plain text.
66
67**Parameters:**
68- `url` (required) — URL to fetch and decompose
69- `compact` (optional, default: false) — Omit zero-value fields
70
71**Example prompt:** "Decompose https://spec.example.com/transport and show me the security requirements"
72
73## What It Detects
74
75- **Authority levels** — RFC 2119 keywords: "shall" = mandatory, "should" = directive, "may" = permissive
76- **Risk categories** — safety-critical, security, compliance, financial, contractual
77- **Attention scoring** — authority weight x risk multiplier, 0-10 scale
78- **Standards references** — ASTM, ASCE, IBC, OSHA, ACI, AISC, AWS, ISO, EN
79- **Financial values** — dollar amounts, percentages, retainage, liquidated damages
80- **Dates** — deadlines, milestones, notice periods
81- **Irreducibility** — legal mandates, threshold values, formulas that cannot be paraphrased
82
83## Use Cases
84
85- Pre-process documents before sending to your LLM — save 60-80% of context window
86- Classify specs, contracts, policies, regulations by obligation level
87- Extract standards references and compliance requirements
88- Route high-attention content to specialized analysis chains
89- Build structured training data from raw documents
90
91## Performance
92
93- ~14ms average per document on Apple Silicon
94- 1,000+ chars/ms throughput
95- Zero API calls, zero cost, works offline
96- Deterministic — same input always produces same output
97
98## Security & Trust
99
100**Text classification is fully local.** The `decompose_text` tool performs all processing in-process with no network I/O. No data leaves your machine.
101
102**URL fetching performs outbound HTTP requests.** The `decompose_url` tool fetches the target URL, which necessarily involves network I/O to the specified host. This is why the skill declares the `network` permission in `claw.json`. If you do not need URL fetching, you can use `decompose_text` exclusively with no network access required.
103
104**SSRF protection.** URL fetching blocks private/internal IP ranges before connecting: `0.0.0.0/8`, `10.0.0.0/8`, `100.64.0.0/10`, `127.0.0.0/8`, `169.254.0.0/16`, `172.16.0.0/12`, `192.168.0.0/16`, `::1/128`, `fc00::/7`, `fe80::/10`. The implementation resolves the hostname via DNS *before* connecting and checks all returned addresses against the blocklist. See [`src/decompose/mcp_server.py` lines 19-49](https://github.com/echology-io/decompose/blob/main/src/decompose/mcp_server.py#L19-L49).
105
106**No API keys or credentials required.** No external services are contacted except when using `decompose_url` to fetch user-specified URLs.
107
108**Source code is fully auditable.** The complete source is published at [github.com/echology-io/decompose](https://github.com/echology-io/decompose). The PyPI package is built from this repo via GitHub Actions ([`publish.yml`](https://github.com/echology-io/decompose/blob/main/.github/workflows/publish.yml)) using PyPI Trusted Publishers (OIDC), so the published artifact is traceable to a specific commit.
109
110## Resources
111
112- [Source Code (GitHub)](https://github.com/echology-io/decompose) — full source, auditable
113- [PyPI](https://pypi.org/project/decompose-mcp/) — published via Trusted Publishers
114- [Documentation](https://echology.io/decompose)
115- [Blog: When Regex Beats an LLM](https://echology.io/blog/regex-beats-llm)
116- [Blog: Why Your Agent Needs a Cognitive Primitive](https://echology.io/blog/cognitive-primitive)
modbender