bdk-lint-tools
Skill1 starsChanged 48 days agoMITView on GitHub
Project-configured lint/format/typecheck commands from .bdk/settings.json. Preloaded into agents that run static analysis; not user-facing.
In this repository26 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- agent-lintSkill
- skill-lintSkill
- add-ruleSkill
- bdk-implementer-return-contractSkill
- bdk-rules-architectureSkill
- bdk-rules-code-qualitySkill
- bdk-rules-design-patternsSkill
- bdk-rules-languagesSkill
- bdk-rules-securitySkill
- bdk-test-toolsSkill
- commitSkill
- create-adrSkill
- create-planSkill
- crSkill
- debugSkill
- designSkill
- explain-complex-codeSkill
- mermaid-drawerSkill
- pr-reviewSkill
- refine-rulesSkill
- setupSkill
- subagent-execute-planSkill
- test-driven-developmentSkill
- update-docsSkill
- verify-planSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
If the line above reads run the project's linter/formatter, .bdk/settings.json is not configured — emit one warning line [bdk] …
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Lint/format/typecheck command(s) for this project, one block per tier:
!python3 ${CLAUDE_PLUGIN_ROOT}/scripts/get_settings.py lint-tools
If the line above reads run the project's linter/formatter, .bdk/settings.json is not configured — emit one warning line [bdk] .bdk/settings.json not configured — run /bdk:setup then detect from project files and proceed under the same policy below. Prefer a project script (bin/cleanup.sh, a Makefile lint target) when one exists and you had to detect.
Pick the form, not just the command
Each block names a tier and the forms of the same command. {files} is a placeholder: replace it with the space-separated paths you were given, quoted.
| Tier | Given a file list | No file list |
lint, format | scoped — these tools take paths natively, so a whole-project sweep on a three-file change is pure waste. | full |
typecheck | incremental — a typechecker resolves the whole program, so a path list buys little; the cache is what makes a repeat run cheap. Fall back to full if no incremental form is configured. | incremental, else full |
The incremental form's cache (.tsbuildinfo and equivalents) survives between runs in the same worktree, so the second and later checks of a run pay only for the delta. Do not delete it between checks.
If a lint/format block has no scoped form, derive one by appending the paths the tool accepts (<full> -- {files} for an npm/yarn/pnpm script, a bare path list for most direct binaries) and say in your report that you derived it, so the settings get fixed once instead of re-derived forever.
Scope policy
- Given file paths, check those paths. Per task, per group, and per fix cycle you get a file list — use it. Findings outside the list are out of scope; do not widen the run to go looking for them.
- Non-executable content gate. If every path you were given is yaml/md/json/plain config not feeding build or codegen: skip
typecheckentirely, and runlint/formatscoped only if the tool supports those file types - otherwise reportnothing to verify for these paths: <list>. Build-feeding config (tsconfig, lockfiles, codegen schemas) counts as source: when the source partition is empty,typecheckmust not run. - The unscoped whole-project sweep belongs to the end-of-run gate, or to a caller who explicitly asked for it.
- Report the exact command you ran (form and substituted paths) with the findings, so a caller can tell what was actually covered.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.