v0.16.3 — sheleg-dev

• "product analytics" / «аналитика продукта» join ad-tracking's ## Triggers (hub board row
  B-267). The skill owns that ground by its own text: its architecture diagram puts Mixpanel
  "(product analytics)" behind the same consent layer as the pixels, and it carries Mixpanel's
  alias / identify rules. The description is unchanged. The body's opening sentence is
  shortened to pay for the words: 4,748 of 4,750 tokens (audit_skill.py --house).

v0.16.2 — sheleg-dev

• Every skill description now fits a budget of about 300 characters (hub board row B-255 (a)):
  Claude Code lists every installed skill's description each session against a shared budget of
  about 30,000 characters, and the operator's listing measured 51,178, so less-used skills lost
  their descriptions. Measured with whitespace collapsed: ad-tracking 408 → 297,
  crypto-payments 398 → 298, error-tracking 395 → 292, frontend-performance 412 → 299,
  google-auth 417 → 293, google-signin 386 → 300, stripe-billing 459 → 299; total 2875 →
  2078.
• Each description keeps its router sentence, its boundary against sibling skills, the opt-out
  phrase and a short English and Russian trigger list. Every quoted trigger the description
  dropped is kept in the body's ## Triggers section, so the family hook's words stay advertised
  (test/advertised_check.js in the hub).
• The CI negative for a duplicated trigger literal now anchors on "google-auth-library", and
  plants two literals google-signin still quotes ("sign in with google", "one tap"): the old
  anchor Triggers - "google auth", left the description. docs/evals/stripe-billing.md carries
  the re-measured body (4647 tokens, 406 lines).

v0.16.1 — sheleg-dev

• test/family_gate.py, the family's canonical member gate (hub B-108, B-109): the board
  parses (every row as wide as its header, no id twice) and release.yml keeps the tag-ancestry
  check (fetch-depth: 0, git merge-base --is-ancestor), each watched failing on a plant by
  --self-test. Wired into npm test; a byte-for-byte copy of the hub's member-gate/family_gate.py.

v0.16.0 — sheleg-dev

• stripe-billing/references/cancellation-and-retention.md gains A pause instead of a
  cancellation: the screen rule (the pause offer does not lengthen the cancel path; one
  screen, cancel one tap away), then Stripe's two pauses side by side. pause_collection
  (behavior keep_as_draft, mark_uncollectible or void; resumes_at) keeps status
active and access on, and reaches the app only as customer.subscription.updated. The
  Pause endpoint (flexible billing_mode, API 2026-09-30.endive) sets paused, stops
  invoices and emits customer.subscription.paused and .resumed.
  status_details.paused.subscription.type separates a requested pause from a trial that
  ended without a payment method (trial_settings.end_behavior.missing_payment_method) and
  from a payment-failure pause. Four test-matrix lines; every fact cites docs.stripe.com,
  read 2026-10-22.
• price-integrity.md cites docs.stripe.com/keys for the key prefixes assertStripeEnv
  reads. Board row B-86 is closed: its assertion and CREDENTIAL_BOUNDARIES row shipped in
  v0.13.4 and the row had stayed open.
• subscription-lifecycle.md: mapStatus maps paused to its own paused, no longer canceled;
  a paused subscription resumes. Board: B-107 (the client's API pin predates the Pause endpoint)
  and B-108 (billing_mode migration) filed.

v0.15.1 — sheleg-dev

• The accessibility reference's image-redundancy example used the name of one of the operator's private projects as its sample brand. It now uses Acme. No behaviour changes.

v0.15.0 — sheleg-dev

• Every skill's front-matter description is now a router, not a manual: one
  "Use when" sentence, a "Not for" boundary naming the sibling skill, the opt-out
  phrase (no wiring, без обвязки) and a short English and Russian trigger list.
  Lengths before and after: stripe-billing 965 to 459, error-tracking 962 to 395,
  frontend-performance 885 to 412, crypto-payments 884 to 398, ad-tracking 859 to 408,
  google-signin 821 to 386, google-auth 798 to 417 (6,174 to 2,875 chars in total).
• Why: Claude Code shares about 30,000 chars of skill listing across every installed
  skill. Measured on 2026-10-20, the operator's listing held 29,995 chars for 134 skills
  and 81 of them showed no description. Plugin skills cannot be hidden, so description
  length is the only lever.
• Every trigger phrase removed from a description moved verbatim into a new
  ## Triggers section at the end of that skill's body, so the family's routing hook
  still finds each of its words in the skill's text.
• test/validate.py refuses a description over 500 chars (the family cap; the
  standard's 1024 stays the hard limit). The CI plant that pads crypto-payments by
  320 chars now trips this cap.
• docs/evals/stripe-billing.md restates the recomputed body budget (4603 tokens,
  403 lines).

v0.14.0 — sheleg-dev

• frontend-performance/SKILL.md gains Test browser — WebKit (Safari) first:
  Safari through safaridriver, Playwright webkit (--browser webkit,
  npx playwright install webkit), Safari Web Inspector's Timelines, and
  open -a Safari <url> on macOS rather than the system default browser.
• Lighthouse, PageSpeed Insights, CrUX, CDP traces and the chrome-devtools MCP stay
  as the stated Chromium exception (Chromium by definition): Chrome Dev, a
  temporary profile, every started process closed. A page checked only in Chromium
  is reported as exactly that; a Lighthouse score is not a cross-browser verdict.
• The Measure and Validate steps run the WebKit pass too; Degradation names the
  WebKit channels; the contrast check in references/accessibility-perf.md starts
  in Safari Web Inspector.
• New regression test/audit_regressions/safari-first.py with drop mutations. Text
  integrity only; no hook or script behaviour changed.

v0.13.4 — sheleg-dev

• Supply a copyable, locally tested Stripe mode assertion before client construction.
• Refuse explicit Google service-account JSON and OAuth client-secret assignments in
  non-production or undeclared runs, while preserving public metadata assignments.
• Correct Google Testing advice for basic identity scopes and Workspace controls.
• Replace a Google access-token log with a value-free event. Document ambient ADC,
  skills-only hooks and CAPI coverage limits. No provider operation was performed.
• Validate explicit repository paths inside skill Markdown without treating consumer
  project paths as repository claims.

v0.13.3 — sheleg-dev

• The release refuses a lightweight tag. release.yml gains one step, right after
  checkout and before any gate: git cat-file -t on the tag must print tag; a commit
  fails the job with the remedy (git tag -a <tag> -m '<release>' at the same commit)
  before anything is published. git describe and git submodule status see annotated
  tags only, and the 2026-10-09 family wave cut four lightweight member tags that the
  umbrella then read as each member's previous release. The step is identical across the
  family (umbrella plan 2026-10-10, T5 / REQ-6). Published tags are not re-cut.
• No skill, reference or installer change.

v0.13.2 — sheleg-dev

Error tracking no longer assumes every non-Claude host lacks MCP or that setup
works only in Claude Code and Cursor. It checks the active host prerequisites,
keeps CLI/SDK fallback, and labels the Claude MCP command as host-specific.
A skills-only installation does not claim to register MCP or hooks.