Run a post-baseline tool-gap and overlap audit

[szmyty]

Outcome

After the current universal MegaLinter remediation pass is complete, perform a deliberate tool-gap and overlap audit before adding any more linters or formatters to EgoLint.

The result should be a small evidence-backed set of adopt/adapt/defer/reject decisions rather than an ever-growing collection of redundant tools.

Context

The current proving-ground work has nearly completed configuration and fixture coverage for the large MegaLinter surface. The next planned step is to ask whether complementary tools add unique value outside that baseline.

Candidate examples surfaced during the work include:

• Flynt for Python modernization transforms;
• shellharden/shfmt for shell formatting or hardening;
• Taplo for TOML;
• dprint for cross-language formatting;
• ty or basedpyright for Python typing experiments;
• other maintained tools discovered during remediation.

Biome is already covered by #13 and Betterleaks has its own evaluation issue; those should be consumed rather than duplicated here.

Evaluation model

Classify capabilities by responsibility:

• formatting;
• linting;
• type checking;
• security/secrets;
• licenses/SBOM;
• spelling/prose/links;
• schemas;
• repository policy;
• CI/workflow validation;
• modernization/refactoring.

For every candidate answer:

1. What concrete gap does it fill?
2. Which existing EgoLint/MegaLinter tool overlaps?
3. Does it replace, complement, or merely duplicate an existing capability?
4. Is it universal, profile-specific, lifecycle-specific, or host/editor-only?
5. Can it be pinned and run deterministically without network installation during lint execution?
6. Does it emit usable machine-readable evidence?
7. Does it support safe check/fix separation?
8. What are maintenance, license, supply-chain, runtime, and image-size costs?
9. Does adding it improve the developer experience enough to justify another dependency/config surface?

Required artifact

Produce a versioned tool matrix with at least:

• tool and version evaluated;
• capability/responsibility;
• applicability predicates;
• canonical owner it would replace or complement;
• overlap score/rationale;
• runtime/image cost;
• editor/hook/CI integration implications;
• output/diagnostic formats;
• fix safety;
• maintenance/license/supply-chain notes;
• decision: adopt, adapt, defer, reject;
• decision rationale and follow-up issue when adopted.

Acceptance criteria

• The current MegaLinter/EgoLint baseline is inventoried before new candidates are judged.
• Every candidate has an explicit gap/overlap analysis.
• Biome work is linked to Add a JavaScript package-quality profile with Oxlint, Biome, and publint #13 rather than redefined.
• Betterleaks work is linked to its dedicated evaluation issue.
• Flynt is explicitly evaluated as modernization/refactoring rather than assumed to be a default linter.
• No tool is adopted solely because it is popular or available.
• Adopted tools have deterministic packaging and applicability plans.
• Rejected/deferred tools retain concise rationale so they are not repeatedly re-investigated without new evidence.
• The final tool matrix feeds Define adaptive capability detection and execution policy #35 capability detection and Define canonical formatter ownership across editors, hooks, tasks, and CI #53 formatter ownership where applicable.

Related

• Add a JavaScript package-quality profile with Oxlint, Biome, and publint #13 — JavaScript package-quality profile with Oxlint/Biome/publint
• Define adaptive capability detection and execution policy #35 — adaptive capability detection and execution policy
• Roadmap: ship the adaptive EgoLint quality distribution #42 — adaptive EgoLint quality distribution roadmap
• Graduate the proven universal MegaLinter policy corpus into EgoLint #51 — graduate the proven universal MegaLinter policy corpus
• Evaluate Betterleaks as the default Git-oriented secret scanner #52 — Betterleaks evaluation
• Define canonical formatter ownership across editors, hooks, tasks, and CI #53 — formatter ownership across editor/hooks/tasks/CI

Non-goals

• Running every available linter or formatter.
• Replacing mature language-native tools without evidence.
• Increasing default image size or runtime merely to maximize tool count.