ai-code-intelligence-toolkit
v1.0.9
Published
AI code intelligence toolkit with anti-drift task contracts, graph queries, TypeDoc source-link support, and agent-safe install injection.
Maintainers
Readme
AI Code Intelligence Toolkit v1.0.8
AI code intelligence toolkit with anti-drift task contracts, graph queries, TypeDoc source-link support, agent-safe install injection, smart validation memory, and durable AI changelog tracking.
Install
npm install --save-dev ai-code-intelligence-toolkit typedoc-hybrid-source-links typedoc
npx typedoc-hybrid-install --target . --overwrite
npx ai-code-intel-install --target . --overwrite --strict
npm run ai:history:initRequired anti-drift edit cycle
npm run typedoc:json:local && npm run ai:graph:build
npm run ai:spec -- "<task>"
npm run ai:preflight -- "<task>"
npm run ai:graph:query -- "<specific symbol/file/error/feature>"The agent may edit every necessary repository file after this discovery flow completes. Preflight and graph query are guidance tools, not a hard edit whitelist.
Token conservation and drift prevention
- Read
AGENTS.md,docs/ai-changelog/START_HERE.md,AI_GROUND_TRUTH.md, andAI_SYMBOL_INDEX.jsonbefore source files. - Use
Select-Stringbounded context in PowerShell instead of broadGet-Contentorrgfirst. - Use
ai:test:smartso a build/test that already passed for the same unchanged fingerprint is not repeated. - Use
ai:history:addso important fixes are recorded as numbered markdown files indocs/ai-changelog/.
New v1.0.8 final commands
npm run ai:history:init
npm run ai:history:status
npm run ai:history:add -- --task "<task>" --summary "<what changed>" --files "file1,file2" --validation "npm run build"
npm run ai:test:status
npm run ai:test:smart -- "npm run test"
npm run ai:test:smart -- "npm run build"
npm run typedoc:strictExample workflow
npm run ai:history:status
npm run typedoc:json:local && npm run ai:graph:build
npm run ai:spec -- "Fix dashboard list overflow and side panel spacing"
npm run ai:preflight -- "Fix dashboard list overflow and side panel spacing"
npm run ai:graph:query -- "dashboard list overflow side panel App"
npm run ai:test:smart -- "npm run build"
npm run ai:history:add -- --task "Fix dashboard list overflow" --summary "Resolved side panel overflow and text clipping" --files "src/App.tsx" --validation "npm run build"Searchable aliases
| Purpose | Command |
|---|---|
| Refresh context | npm run ai:context:refresh |
| Find context | npm run ai:context:find -- "<query>" |
| Spec | npm run ai:task:spec -- "<task>" |
| Preflight | npm run ai:task:preflight -- "<task>" |
| Durable history | npm run ai:history:status |
| Smart validation | npm run ai:test:smart -- "npm run build" |
| Final health | npm run ai:health:final |
Companion package
Install with TypeDoc Hybrid Source Links for the complete tested workflow:
npm install --save-dev typedoc-hybrid-source-links ai-code-intelligence-toolkit typedocCumulative v1.0.8 note
This package is the complete v1.0.8 jump release. It carries the package manifest repair, anti-drift agent contract, AGENTS/README injector, edit-permission fix, searchable aliases, durable AI changelog memory, and smart validation memory in one release.
