Every PR that changes product-user-visible GUI behavior now includes a
demonstration GIF with real provenance (that branch's built tree, real
key, real model rounds). Recording stays side-effect-free; the skill
gains a bounded final publication step: GIFs go on an append-only
orphan assets branch (one per PR series) and embed via the blob URL
with ?raw=true, never on the PR branch itself.
Folds in the operational lessons from the Code Mode UI series:
.playwright-mcp/ screenshot roots (now gitignored), per-PR staging and
precise server teardown, one-call DOM polling for transient states,
exact-text completion predicates, prompt engineering for UI states,
and the export-before-invoke GIF_SKILL_DIR encoder pitfall.
Agent Note: implemented/process/2026-07-26-gui-pr-gif-evidence-and-assets-branch (+ zh pair); the 2026-07-23 recording note now defers publication policy to it.
Responding to review on #661: a public method on the generic ToolRegistry
service whose only caller is the run_code bridge was ad-hoc surface
widening. The bridge now receives it as a registry-private capability
closure in RunCodeBridgeOptions (the requireRuntime idiom, alongside the
cap), the method is private, and it leaves the generated service
catalog/API surfaces.
The pattern is now named as a code smell where reviewers look: the
packages/AGENTS.md capability-interface rule gains the inverse-smell
clause (ceiling 660→675 — the list is at capacity and the clause needs
one sentence), and dsh-code-review's capability-fit check tells reviewers
to flag single-consumer public service methods and require the closure
form.
A repo-wide Not Invented Here audit (ten parallel domain surveys covering
every package group, scripts/, native/, vendor/ edges, python/, test
infrastructure, and CI) asked of each hand-rolled surface whether a
maintained external package or Node builtin deletes it with a net win.
Policy: new implemented process note records that a dependency which
genuinely deletes owned code is a preferred simplification (bar: net
deletion, health, boundary fit, settled seams stay); root AGENTS.md
carries the one-line rule and dsh-find-simplifications now surveys for
hand-rolled-where-a-dependency-exists candidates.
Findings, all bilingual from birth:
- proposed/simplification: eventsource-parser for llm-deepseek SSE,
node:timers/promises for three hand-rolled sleeps, turndown (or
minimal 'entities') for tool-web HTML->markdown, gate-script
consolidation onto mdast/parseArgs/globSync
- proposed/testing: execa + parseArgs + loadEnvFile + vi.waitFor for
hand-rolled test subprocess plumbing
- proposed/process: pnpm/action-setup for symmetric CI caching
- proposed/feature: evaluate landstrip before building a Windows
sandbox launcher
- rejected/simplification: ~30 swap verdicts recorded (vscode-jsonrpc,
p-retry, Ajv, write-file-atomic, msw, hono, better-sqlite3, wireit,
landstrip-for-linux, YAML consolidation, ...) so the survey is not
re-litigated from scratch
Also drops the stale prompt/ entry from the AGENTS.md layout map
(workspace instructions live in packages/context/workspace-context).
A source-of-truth link did not make the review obligation explicit, so reviewers could treat green documentation checks as sufficient. Make dsh-prose-standard a blocking requirement for every added or changed prose form and require semantic comparison with the owning code or behavior.
A saved complete-file candidate could overwrite newer skill guidance, while a provider-wide outage was indistinguishable from a healthy no-op run. Record the source blob and provenance, reject drift during promotion, and make total adapter failure visible.
Align cadence and feedback acquisition with the evidence contract, and narrow the borrowed-state and negative-control checks so they do not contradict the package and testing standards.
Tianyi Cui