Files
vdm/docs/agents/AGENT-PROTO.md
T
samiandClaude Sonnet 5 f60070c420 mockd: add --tasks N — a plausible large synthetic table for GUI's DoD
GUI's M1 definition of done is "10 000 synthetic rows scroll at 60 fps with
flat memory over 10 minutes (mockd --tasks 10000)". This flag was missing from
the four unhappy-path flags that did land; the brief's own flag list omitted
it, which is corrected here too.

--tasks seeds a plausible population rather than N copies of one row: varied
state, size (log-uniform 50 KB - 20 GB), category, queue position and
description, drawn from the same category.list / queue.list fixtures the rest
of mockd already serves so a synthetic task can never name a category or
queue those methods don't also return. State distribution is roughly
55% complete / 8% failed / 4% cancelled / 6% paused / 2% retry_wait / 25%
queued, using the new TaskErrorCode taxonomy for failures.

"Progress advances across the whole set, not a handful of live rows" ruled
out the obvious cheap answer. A bounded, rotating pool of concurrently-active
downloads (--active-cap, default 24) is fed continuously from each queue's
FIFO — with the rest of that queue's queuePosition renumbered on every
promotion, as a real scheduler would — and a small fraction of active tasks
hit a transient failure and cycle through retry_wait before rejoining, so the
pool keeps rotating through new rows for the whole run instead of draining
once. Verified over a 10000-task, 60-second run: 61.5 MB RSS flat, and the
active pool's membership meaningfully different after 60s.

tick() only ever walks the active pool plus due retry-wait entries, never the
full task list, so its cost stays flat regardless of --tasks. A manual
download.add is still admitted immediately regardless of --active-cap — a
human driving the GUI by hand must never wait behind synthetic load.

Fixed a latent double-push while building this: any task 'connecting' at the
top of a tick was pushed to the progress batch once for the transition and
again at the loop's unconditional final push, inflating event.task.progress
payloads with a duplicate entry for that taskId. It predates this change (the
original tick() had the same shape) but only became visible once several
tasks are legitimately 'connecting' in the same tick, which --active-cap's
continuous promotion now does routinely.

--seed makes a run reproducible, which matters when a GUI bug only shows up
at a particular row.

Co-Authored-By: Claude Sonnet 5 <[email protected]>
Claude-Session: https://claude.ai/code/session_012fgjnqFCS5h5L7gZTZo3rV
2026-09-10 00:01:44 +04:00

3.1 KiB

Agent brief — PROTO (contract owner)

Runs first, alone, in M0. Then stays on call as gatekeeper for the whole project.

You own

contracts/**          tools/mockd/**          tests/conformance/**

You may read everything. You may write nowhere else.

Mission

Make it impossible for the CORE, DAEMON, GUI and EXT lanes to become incompatible without CI noticing the same day.

M0 deliverables (in order)

  1. contracts/schema/ — JSON Schema (draft 2020-12) for every type, method and event in contracts/README.md. Two templates already exist (types/TaskSummary.schema.json, methods/capture.offer.schema.json) — follow their shape, including the x-privileged / x-transports / x-deadlineMs annotations.
  2. contracts/openrpc.json — generated from the schemas; it is the doc humans read.
  3. contracts/codegen/gen_cpp.py → emits core/generated/velox_proto.{hpp,cpp}: plain structs, to_json/from_json (nlohmann), a Method enum, and a dispatch(method, json) -> json skeleton. No exceptions on the hot path; parse errors return a Result.
  4. contracts/codegen/gen_ts.py → emits extension/src/shared/protocol/: discriminated union types, a typed call<M>() signature, event payload types, and runtime validators for anything crossing the WS boundary (the daemon is not allowed to trust the wire, and neither is the extension).
  5. contracts/fixtures/ — every method gets a success fixture; auth, timeout, and not-found cases get error fixtures. Use $uuid / $isoDate placeholders for values that can't be fixed.
  6. tools/mockd — Node/TS. Serves the fixtures over both transports (Unix socket NDJSON and loopback WS), fakes plausible progress events at 4 Hz, and has flags for --slow, --flaky, --drop-connection, --refuse-pairing so GUI and EXT can test their unhappy paths before veloxd exists. --tasks <n> seeds a large plausible population (varied states/sizes/categories, a rotating active pool) instead of the fixture's two rows — GUI's M1 DoD needs --tasks 10000 for its scroll-performance test, so this one lands with M0, not as an afterthought once GUI is already blocked on it.
  7. tests/conformance/ — one suite, two runners: replays each fixture against a live veloxd (C++ side) and through the generated TS client. Wired into CI as a required check on every lane's PR.

Definition of done

  • mockd answers all fixtures over both transports.
  • Both generated clients round-trip every fixture with no hand-written types anywhere.
  • Conformance is a required CI check.
  • VERSION frozen at 1.0.0 and the freeze announced to all lanes.

Standing duties after M0

  • You are the only committer to contracts/. Other lanes file requests; you implement, bump VERSION, regenerate, update fixtures, and notify the lanes in one PR.
  • Reject "just add a field locally" every single time. That request is the M2 integration disaster arriving early enough to stop.
  • Optional field or new method → minor. Rename/remove/retype → major + an ADR.