feat: Adds typescript domain - /tsc-blindspots, /derive-types, /avoid-any, /decompose-large-files, /migration-context-cost skills

[MajorLift]

Overview

Adds a typescript domain — five experimental skills for TypeScript authoring and JS→TS migration, the layer coding-guidelines only gestures at. Grounded in MetaMask/contributor-docs docs/typescript.md.

• avoid-any — any disables type checking rather than widening a type. Substitute by position: assignee → unknown, assigned → never.
• derive-types — derive from authoritative sources (indexed access, typeof) instead of ad-hoc types that duplicate and drift (#42583).
• decompose-large-files — split by coherent, independently-mergeable units, unblocking incremental migration. Modeled on the MetamaskController decomposition (#41735).
• migration-context-cost — the fan-in/fan-out cost of a conversion: cheap first migration, or one pulling half the repo.
• tsc-blindspots — the defects tsc is structurally unable to report; a green build is not evidence the types are correct. Lands here rather than in testing, superseding Add typescript-compiler-blindspots skill #85: derive-types' question, one step later.

Motivation

All three repos author new TypeScript and re-enable @typescript-eslint/no-explicit-any to error over a shared config that turns it off — extension .eslintrc.js:162, mobile .eslintrc.js:170, core eslint.config.mjs:152.

With checkJs off, unconverted neighbors are unchecked, so a green build does not establish well-formedness across the boundary.

Two are proposed always-loaded, spending shared description budget: tsc-blindspots, because a green build reads as proof, so nobody searches for what tsc cannot report; derive-types, because an author hand-writing a type does not know an authoritative source exists.

Showcase

Trial runs against merged PRs nobody flagged. Gap: the run surfaced something the PR had not.

PR	Skill	Verdict	Finding
#41333	tsc-blindspots	Gap	local type restates upstream Log.topics; naive probe reports false clean

The evidence showcase carries the derive-types example: #44397 (derived-type substitution) — 12 hand-written types, 9 with an authoritative source, 5 of those disagreeing, 1 dropped guard traced to inert, 2 unnecessary runtime changes; one arm silent at exit 0.

[MajorLift]

Context budget

What this PR costs an agent, measured from an install rather than read from the diff. Three tiers, and only the first is unavoidable.

Skill	Frontmatter	Selected	+ refs & knowledge
avoid-any	359 chars	~1,605 tok	~1,605 tok
decompose-large-files	266 chars	~1,612 tok	~1,612 tok
derive-types	210 chars	~1,168 tok	~1,168 tok
migration-context-cost	243 chars	~781 tok	~781 tok
tsc-blindspots	1,244 chars	~3,406 tok	~8,327 tok

Frontmatter is the only tier paid unconditionally — every agent loads it on every run once the skill is installed, used or not, because it is what the agent reads to decide relevance. The 28 skills across the eleven open skill PRs sit at a median of ~1,716 tokens selected and ~1,860 with references followed. All are within the 1,536-character description budget.

Selected is paid only when the agent picks the skill. + refs & knowledge is the ceiling if every bundled reference is then read; it is a worst case, not an expectation.

Method

tools/install --repo metamask-extension --maturity experimental against this branch at e170b394c, measured per installed skill directory. Repo overlays are merged into the emitted SKILL.md, so they land in the selected tier rather than being missed by a source-byte count. Token figures are bytes/4 — a proxy for scale, not accounting.

These figures are pinned to the commit above and drift on every push; #96 tracks automating them.