Exposes 11 tools that look up structured failure knowledge across 2,000+ code errors and country-specific dead ends. You get back regex-matched patterns showing what not to try (with failure rates), what works (with success rates), and error chains that predict downstream failures. Useful when you want an agent to skip known-bad fixes like sudo pip or avoid cultural missteps like tipping in Tokyo. Also includes batch lookup, domain stats, and outcome reporting to feed real results back into the dataset. The full canon pulls from government sources for country rules and community validation for code fixes. Python SDK available, or call the JSON API directly at deadends.dev/api/v1.
Public tool metadata for what this MCP can expose to an agent.
lookup_errorMatch an error message against deadends.dev's database of known errors. Returns dead ends (what NOT to try), workarounds (what works), and error chains (what comes next).1 paramsMatch an error message against deadends.dev's database of known errors. Returns dead ends (what NOT to try), workarounds (what works), and error chains (what comes next).
error_messagestringget_error_detailGet full details for a specific error by its ID (e.g., 'python/modulenotfounderror/py311-linux').1 paramsGet full details for a specific error by its ID (e.g., 'python/modulenotfounderror/py311-linux').
error_idstringlist_error_domainsList all error domains and counts in the deadends.dev database.1 paramsList all error domains and counts in the deadends.dev database.
sort_bystringsearch_errorsSearch errors by keyword across all domains. Uses fuzzy keyword search.3 paramsSearch errors by keyword across all domains. Uses fuzzy keyword search.
limitintegerquerystringdomainstringlist_errors_by_domainList all errors in a specific domain with their fix rates.2 paramsList all errors in a specific domain with their fix rates.
domainstringsort_bystringbatch_lookupLook up multiple error messages at once. Returns the best match for each error.1 paramsLook up multiple error messages at once. Returns the best match for each error.
error_messagesarrayget_domain_statsGet detailed statistics for a domain: error counts, average fix rate, resolvability breakdown.1 paramsGet detailed statistics for a domain: error counts, average fix rate, resolvability breakdown.
domainstringget_error_chainTraverse the error transition graph for a specific error. Shows leads_to, preceded_by, and frequently_confused_with.1 paramsTraverse the error transition graph for a specific error. Shows leads_to, preceded_by, and frequently_confused_with.
error_idstringStop AI agents from repeating known failures - in code AND in the real world.
AI assistants reliably fumble two kinds of problems: known-failed code fixes, and country-specific real-world rules they've never been exposed to in training. deadends.dev now covers both:
ModuleNotFoundError, CUDA OOM, CrashLoopBackOff, etc.Why the expansion? Coding dead ends are largely solved by a good LLM. Country-specific friction - Japanese hanko requirements, Schengen 90/180 math, Ramadan business hours, Saudi alcohol ban, Indian beef taboos - is where generic AI advice breaks hardest. The codebase and schema are identical; the env segment just carries a country code.
90% Precision@1 · 0.935 MRR · Data Quality Dashboard
Website: deadends.dev · MCP Server: Smithery · PyPI: deadends-dev · API: /api/v1/index.json Repository: https://github.com/dbwls99706/deadends.dev
| Without deadends.dev | With deadends.dev |
|---|---|
Agent tries sudo pip install → breaks system Python → wastes 3 retries | Agent sees "dead end: sudo pip - fails 70%" → skips it immediately |
| Agent tells user to tip 15% at a Tokyo restaurant | Agent knows tipping is refused in Japan (culture/tipping-refused/jp) |
| Agent drafts a Thai social post referencing King Rama X | Agent stops: Article 112 lèse-majesté risk (legal/lese-majeste-article-112/th) |
| Agent fixes error A, gets confused by error B | Agent knows "A leads to B 78% of the time" → handles both |
| Agent tells unmarried couple to kiss publicly in Dubai | Agent flags UAE public decency law (legal/unmarried-public-affection/ae) |
What makes this different from asking an LLM?
{domain}/{slug}/{env} - env holds the country
code (kr, jp, us, de...) so the same taboo can be answered
differently for different jurisdictions.report_outcome로 빠르게 보완합니다.pip install deadends-dev
deadends "CUDA error: out of memory"
Add to ~/.claude/claude_desktop_config.json:
{
"mcpServers": {
"deadend": {
"command": "python",
"args": ["-m", "mcp.server"],
"cwd": "/path/to/deadends.dev"
}
}
}
Or install via Smithery (no local setup):
npx -y @smithery/cli@latest install deadend/deadends-dev --client claude
Unauthorized 빠른 해결 가이드 (사람용)deadend: calling "initialize": sending "initialize": Unauthorized 에러가 보이면 아래를 순서대로 그대로 실행/확인하세요.
# 로컬 서버 확인 (정상 시 툴 목록이 출력됨)
python -m mcp.server --help
cwd는 실제 경로여야 함)cat ~/.claude/claude_desktop_config.json
cd /path/to/deadends.dev
python -m mcp.server
npx -y @smithery/cli@latest uninstall deadend/deadends-dev --client claude
npx -y @smithery/cli@latest install deadend/deadends-dev --client claude
# macOS 예시
osascript -e 'quit app "Claude"'
open -a Claude
팁:
Unauthorized는 보통 잘못된cwd, 중복 서버 설정(로컬+원격 동시), 또는 만료된 인증 상태에서 발생합니다.
Add as a remote MCP server - no authentication required:
{
"mcpServers": {
"deadend": {
"serverUrl": "https://deadends.dev/mcp",
"type": "http"
}
}
}
Note: Antigravity uses
serverUrl(noturl). If you getUnauthorized, remove any existing deadend entries from the MCP Store and re-add manually using the config above. See the Antigravity MCP auth guide for general troubleshooting.
from generator.lookup import lookup, batch_lookup, search
# Single error lookup
result = lookup("ModuleNotFoundError: No module named 'torch'")
# What NOT to try (saves tokens and time)
for d in result["dead_ends"]:
print(f"AVOID: {d['action']} - fails {int(d['fail_rate']*100)}%")
# What actually works
for w in result["workarounds"]:
print(f"TRY: {w['action']} - works {int(w['success_rate']*100)}%")
# Batch lookup (multiple errors at once)
results = batch_lookup(["error1", "error2", "error3"])
## ModuleNotFoundError: No module named 'X' (Python 3.11+)
Resolvable: true | Fix rate: 0.88
### Dead Ends (DO NOT TRY):
- pip install X with system Python (fails 70%): venv not activated
### Workarounds (TRY THESE):
- Create venv, activate, then pip install (works 95%)
- Use python -m pip install instead of bare pip (works 90%)
| Tool | Description |
|---|---|
lookup_error | Match an error message against 2000+ known patterns |
get_error_detail | Full canon by ID |
list_error_domains | All 54 domains with counts |
search_errors | TF-IDF keyword search across all domains |
list_errors_by_domain | All errors in a domain |
list_errors_by_country | All country-scoped dead ends for an ISO alpha-2 code |
get_country_summary | Country-level summary (entries, fix rate, domain mix) |
batch_lookup | Look up multiple errors at once (max 10) |
get_domain_stats | Domain quality metrics and confidence levels |
get_error_chain | Traverse the error transition graph |
report_outcome | Report whether a workaround worked (feeds back into success rates) |
| Endpoint | Description |
|---|---|
/api/v1/match.json | Lightweight regex matching (fits in context window) |
/api/v1/index.json | Full error index with metadata (entries include country field for country canons) |
/api/v1/{id}.json | Individual ErrorCanon |
/api/v1/countries.json | Country index with counts and update dates |
/api/v1/country/{cc}.json | Per-country aggregate (one call returns all entries for that country) |
/api/v1/openapi.json | OpenAPI 3.1 spec |
/api/v1/stats.json | Dataset quality metrics by domain |
/api/v1/errors.ndjson | NDJSON streaming |
/llms.txt | LLM-optimized listing (llmstxt.org) |
/dashboard/ | Data quality dashboard |
| Domain | Errors | Examples |
|---|---|---|
| Python | 88 | ModuleNotFoundError, TypeError, KeyError, MemoryError, RecursionError |
| Node | 70 | ERR_MODULE_NOT_FOUND, EACCES, EADDRINUSE, heap OOM, ERR_REQUIRE_ESM |
| Docker | 65 | no space left, exec format error, bind address in use, healthcheck |
| Kubernetes | 61 | CrashLoopBackOff, ImagePullBackOff, OOMKilled, RBAC forbidden, HPA |
| Git | 60 | failed to push, merge conflicts, detached HEAD, stash apply, tags |
| CUDA | 57 | OOM, device-side assert, NCCL, cuDNN, tensor device mismatch |
| Go | 54 | nil pointer, unused import, interface conversion, slice out of range |
| Java | 54 | NullPointerException, ClassNotFound, OutOfMemoryError, connection pool |
| Database | 52 | deadlock, connection pool, slow query, replication lag |
| AWS | 51 | AccessDenied, S3 NoSuchBucket, Lambda timeout, CloudFormation rollback |
| .NET | 50 | NullReferenceException, LINQ translation, DI circular, EF concurrency |
| ROS 2 | 50 | node spin, launch error, QoS mismatch, tf2 transform |
| TypeScript | 49 | TS2307, TS2322, TS2345, TS2532, TS7053 |
| Rust | 48 | E0382 borrow, E0308 mismatch, E0277 trait, E0106 lifetime |
| + 37 more domains | 40+ each | CI/CD, PHP, Terraform, Networking, Next.js, React, pip, Android, ... |
| Domain | Covers | Example dead ends |
|---|---|---|
visa | Pre-travel authorization, overstay, re-entry bans | ESTA 90-day rule (US), K-ETA (KR), ETIAS/EES (Schengen), Schengen 90/180 (DE) |
banking | Account opening, KYC, foreigner rules | ARC required (KR), residence card 6-month (JP), SSN/ITIN (US) |
emergency | Correct emergency numbers, transit | 112 not 911 (DE), 999/101/111 (UK) |
medical | Insurance, Rx import, coverage | Shaho/Kokuho (JP), NHIS 6-month (KR), EHIC ineligibility (DE), Adderall import ban (JP) |
legal | Criminal liability, contract norms | §86a Nazi symbols (DE), Article 112 (TH), Article 301 (TR), alcohol ban (SA), key money (JP) |
culture | Etiquette, taboos, social norms | Chopsticks in rice (JP), clock gifts (CN), Tiananmen silence, red ink names (KR), bonjour (FR) |
food-safety | Water, pathogens, religious taboos | Tap water (MX), fugu license (JP), beef in India, pork in Indonesia |
communication | Language register, terminology | Honorifics (KR), American War framing (VN), 'gringo' (MX), Cantonese vs Mandarin (HK) |
safety | Driving, public-safety norms | Left-side drive (JP), Autobahn rules (DE), horn-language (IN) |
All metrics are publicly available on the Data Quality Dashboard:
report_outcome updates fix success rates from real usagereview_status: human_reviewed), no LLM bulk generationkr · jp · us · de · uk · fr · it · es · nl · ch · pt ·
ie · at · be · se · no · dk · fi · pl · gr · cn · hk ·
tw · th · in · vn · id · sg · ph · my · pk · bd · sa ·
ae · tr · il · ru · br · mx · ar · cl · co · pe · au ·
nz · eg · ma · et · ng · ke · za · ca
See /country/ hub or /api/v1/countries.json for the authoritative list with counts.
See docs/country-canon-guide.md for the
authoring workflow, sourcing requirements, and confidence calibration.
See CONTRIBUTING.md for full details.
GitHub 자동 수집 운영안: docs/GITHUB_DATA_COLLECTION_STRATEGY.md
자동 수집 주기: 6시간마다(하루 4회), 기본 품질 필터: min_score=2
수집 데이터는 후보이며, 최종 반영은 maintainer 검수 후 진행
Use report_outcome via MCP after trying a workaround
pip install -e ".[dev]"
python -m generator.pipeline # Full pipeline
python -m generator.build_site # Build static site
python -m generator.validate # Validate data + site
python -m pytest tests/ -v # Run tests
ruff check generator/ tests/ # Lint
python benchmarks/run_benchmark.py # Run benchmarks
아래 명령은 템플릿에 핵심 SEO 신호가 있는지 빠르게 점검합니다.
python - <<'PY'
from pathlib import Path
files=[
'generator/templates/index.html',
'generator/templates/domain.html',
'generator/templates/error_summary.html',
'generator/templates/page.html',
'generator/templates/search.html',
'generator/templates/dashboard.html',
]
required=[
'<title',
'meta name="description"',
'meta name="robots"',
'link rel="canonical"',
'meta property="og:title"',
'meta name="twitter:card"',
]
for f in files:
txt=Path(f).read_text()
missing=[r for r in required if r not in txt]
print(f'✅ {f}' if not missing else f'❌ {f} missing: {", ".join(missing)}')
PY
실제 빌드 결과물까지 확인하려면:
python -m generator.build_site
python -m http.server -d public 8080
그 후 브라우저에서 아래를 점검:
view-source:http://localhost:8080/search/view-source:http://localhost:8080/dashboard/<priority> weighted by page strength (country/high-evidence
pages ranked higher); JSON-LD omits empty datesvisa, banking, emergency (plus extended use of
existing legal, culture, medical, communication, food-safety,
safety domains with country env segment)/country/{cc}/ (e.g.
/country/jp/)generator.country_canon_template helper for authoring new country
canons with validated env-segment + audience + jurisdiction metadatadocs/country-canon-guide.md: sourcing standards (primary > embassy
reputable media), confidence calibration, slug/regex conventions
/dashboard/ - transparent metricsreport_outcome → aggregated stats → fix_success_rate updatesreport_outcome)page_url field added to index.json, errors.ndjson, and all SDK/MCP responsesMIT (code) · CC BY 4.0 (data)
docs/SEO_OPERATIONS_GUIDE.mddocs/PYPI_RELEASE_MANUAL.mdcsoai-org/pdf-document-mcp
xt765/mcp-document-converter
io.github.xjtlumedia/markdown-formatter
io.github.ai-aviate/better-notion
suekou/mcp-notion-server
meterlong/mcp-doc