refactor: hard-fork GSD -> MSD (Make Software Done)

Mechanical rename produced by scripts/msd-rename.cjs: gsd/Gsd/GSD -> msd/Msd/MSD
across contents and paths, upstream package/repo coordinates -> @golem15/msd-core
and golem15com/msd-core. Deep links into upstream history, sibling upstream
packages, the GSD-2 import feature, CHANGELOG.md and .changeset/ are kept as-is.

Hand edits on top: MSD block-letter banner and logos, LICENSE copyright line,
package/plugin identity, regenerated lockfile, install-tree fixtures, derived
registries and benchmark baseline; migration checksum baseline re-locked
(MSD keeps its own install state, so no install had applied the old sums);
sort-order and regex-escaped expectations in tests adjusted.
This commit is contained in:
Jakub Zych
2026-10-06 01:47:40 +02:00
parent fe069b2a56
commit a9a7a328e6
2763 changed files with 78465 additions and 78434 deletions

View File

@@ -0,0 +1,138 @@
# Loop Hook Dispatch Contract
Generic reference for consuming the `--raw` JSON output of `msd_run loop render-hooks <point>`
in any host-loop workflow. This document is point-agnostic — it applies to every loop
extension point (discuss:pre, discuss:post, plan:pre, plan:post, execute:pre, execute:wave:pre,
execute:wave:post, execute:post, verify:pre, verify:post, ship:pre, ship:post).
## Envelope shape
```json
{
"point": "discuss:pre",
"activeHooks": [
{ "kind": "contribution", "into": "orchestrator", "fragment": { "inline": "..." } },
{ "kind": "step", "ref": { "skill": "my-skill" } },
{ "kind": "gate", "check": { "query": "..." }, "blocking": true, "onError": "skip" }
],
"rendered": "..."
}
```
`activeHooks` is an array of enabled hook entries for the named point. It is empty (or absent)
when no capability has registered an active hook at this point — treat that as a no-op.
## Dispatch rules by `kind`
### `contribution`
Inject `fragment.inline` verbatim into the context for the role named in `into`
(e.g. `orchestrator`, `planner`). Do not paraphrase — the text is the product.
### `step`
Dispatch the referenced unit. Exactly one of `ref.skill`, `ref.agent`, or `ref.command` is set.
- `ref.skill` present → dispatch via the Skill tool with skill id `msd-<ref.skill>`.
- `ref.agent` present → dispatch via the Agent tool with `subagent_type` = `ref.agent`.
Before dispatching an agent, print the canonical liveness banner so users know silence
is expected and do not kill a healthy agent:
```
◆ Spawning <agent>... (runs in a subagent — no output until it returns; expected, not a freeze)
```
- `ref.command` present → **validate it IN-CONTEXT first, before any shell use.** It comes
from a capability manifest, which may be third-party. Check the value you read from
`activeHooks` against `^[a-z][a-z0-9-]*( [a-z][a-z0-9-]*)*$` yourself — **never** by pasting
it into a shell command to be tested there, because a value carrying a quote, `;`,
`` ` ``, `$(`, or a newline would terminate the assignment and run as its own statement
before any shell-side check could execute. A value that fails is a malformed manifest:
record a warning, skip that hook, continue to the next entry. Only a value that has passed
is run, with the phase number appended:
```bash
msd_run ${ref.command} --phase "${PHASE_NUMBER}" --raw
```
Wait for the result before continuing to the next hook or the next step.
**`supportsReviewerLanes` (optional, boolean).** A `step` entry may carry
`supportsReviewerLanes: true` alongside `ref` (#4209). A workflow opts a step into external
reviewer-lane dispatch by calling `msd_run review-lane dispatch-step --cap-id <capId> --point
<point> --explicit <slugs> ...` — `dispatch-step` resolves its OWN active hook for `<point>` (via
`resolveActiveHooksForPoint`, the same in-process resolver `loop render-hooks` itself calls) and
checks whether `<capId>`'s hook carries this field before proceeding; the workflow does not
resolve or gate on the trait itself, only passes the two flags naming which step it is. When the
trait reads exactly `true`, `dispatch-step` routes through `dispatchReviewerLanes`, the one
interpreter in `src/reviewer-step-dispatch.cts` that reuses the existing reviewer-lane selection,
planning, and invocation machinery, so any explicitly selected reviewer lane also reviews the
same scope. Absent or `false` is inert: `dispatch-step` itself is a no-op (a non-boolean value is
rejected by `capability-validator.cjs` at load time, so it never reaches `dispatch-step` at all).
This is the only place a step opts into reviewer-lane support: a capability beyond `code-review`
reuses it by declaring the same trait on its own step and calling `dispatch-step` with
`--cap-id`/`--point`, with zero bespoke TRAIT-RESOLUTION code of its own. The workflow still owns
matching its own CLI flags against the reviewer-lane roster and assembling the evidence block
handed to its consolidator — those are NOT part of what this trait makes reusable.
A `step` is **advisory by construction**: it never blocks or redirects the host workflow —
that is what a `gate` is for. Each dispatch is best-effort; on error record a warning and
continue, honoring `onError`.
**A point whose workflow hand-rolls one `kind` does not implement this contract.** Several
host workflows historically matched a single hook (e.g. `execute:post` matched only
`ref.skill == "code-review"`), so any other step registered there was declared and silently
never run. When a workflow defers to this file, it dispatches **every** active `step` entry,
not one shape of one.
### `gate`
**Validate `check` before any shell use.** `check.query` and `check.predicate` come from a
capability manifest, which may be third-party — and `gates[].check` is **not** one of the
executable surfaces the install consent prompt discloses (`hooks`, command modules,
`mcpServers`, reviewer lanes), so a capability can be consented to as declarative-only and
still reach a shell through a gate. Check the query value you read from `activeHooks` against
`^[a-z][a-z0-9-]*( [a-z][a-z0-9-]*)*$` yourself, IN-CONTEXT — **never** by pasting it into a
shell command to be tested there, because a value carrying a quote, `;`, a `` ` ``, `$(`, or a
newline would terminate the assignment and run as its own statement before any shell-side
check could execute. A value that fails is a malformed manifest: record a warning, route it
per the gate's `onError`, and do not run it. Pass `check.predicate` as a **single argv
element** for the same reason — never re-quote it into a shell string, where an apostrophe
would close the literal. This is the identical requirement `step` → `ref.command` carries
above; it was stated there and omitted here, which is the gap #3559 closed.
Evaluate `check` (one of `query`, `predicate`, or `agentVerdict`). Then honor `blocking`:
- `blocking: true` → if the check returns `block: true`, surface `check.message` to the user
and stop the current step. Do not continue.
- `blocking: false` → advisory only; surface the message but continue regardless of outcome.
Honor `onError` if the check itself errors: `skip` means treat as non-blocking and continue;
`halt` means surface the error and stop.
## Empty / absent `activeHooks`
If `activeHooks` is absent, null, or an empty array, skip silently and continue to the next
step in the workflow. No output to the user is needed.
## The `execute:task` point (a different shape)
`execute:task` exists below wave granularity — it is evaluated once per task, inside the
`execute:wave:pre` / `execute:wave:post` bracket, immediately before that task's `read_first`
gate. It is **not** one of the 12 points documented above, does not appear in `steps` /
`contributions` / `gates`, and is never dispatched through `msd_run loop render-hooks <point>` or
this file's `activeHooks` envelope.
Instead, a capability declares task-content resolution directly in its manifest body via
`taskContentResolver` (`trackerPrefix` + a bounded `invoke`) — see
[Capability manifest reference](../../docs/reference/capability-manifest.md). `execute-plan.md`'s
per-task loop calls `msd_run task resolve-content --plan <path> --task-id <tracker-id> --raw`
directly, an unconditional, required subprocess invocation with a real, binding exit code —
never a prose-dispatched `step`/`gate` entry chosen from an `activeHooks` array.
This point always runs — there is no `when` config gate and no autonomous-mode elision. That is
deliberate, not an oversight: the twelve points above are best-effort prose dispatch, which
`execute:task`'s hard-halt safety property cannot be built on top of (a missed dispatch is
indistinguishable from a legitimate resolver-empty fallback). See
[ADR-3646](../../docs/adr/3646-per-task-content-resolution-seam.md) for the full rationale,
including why a `kind: "gate"` shape was rejected outright.