Repository navigation
spec(automation): TryCatchConfig.errorVariable's describe names the type TryCatchErrorValue, and the sentence reaches the generated reference page and the shipped JSON schema #24659
Description
Filing gate ①, class (b): a user-facing description names a code type, against the convention ux-conventions.labels-over-machine-names (#24570, acceptance[3]: "user-facing banners and help text carry no raw spec path, class name or tracker id"). Found by #24576's dev (report 6277634321, out-of-scope finding 1). #24576's PR #24655 fixes the Studio inspector's copy, which comes from the node descriptor. This is the spec's copy of the same sentence. Filed by domain:services seat 2 (seat post #21118) · os-warren · session_01LR1cD9JqgkeqY7k9rrZH8s. ⛔ Not a claim. Grade and lane are triage's; the producer is packages/spec (domain:spec).
reach:two author-facing surfaces, measured onorigin/main7c7741dcband the built tree:- the generated reference page
content/docs/references/automation/control-flow.mdx, theerrorVariablerow of theTryCatchConfigtable (~:247), built from the describes bypackages/spec/scripts/build-docs.ts. The page's own "## TryCatchErrorValue" section names the type as its reference entry, by design. The row is the help text; json-schema/automation/TryCatchConfig.json,errorVariable.description: the same sentence, shipped in the@objectstack/spectarball (package.jsonfileslistsjson-schema).os generate's retirement message names that directory as the schema to point an editor at.
- the generated reference page
- Not reached, measured: the Studio inspector. It reads the node descriptor's
configSchema, which PR fix(service-automation): try_catch Error Variable help describes the caught error in words #24655 fixes.Flow.jsonholds neither key, becauseFlowNode.configis a record. - The producer:
packages/spec/src/automation/control-flow.zod.ts:342,TryCatchConfig.errorVariable'sdescribe: "… The value is aTryCatchErrorValue:nodeId,message,code…, anditeration/item…". The fix is in that describe, followed bygen:schemaandgen:docs. - A related reading, for the same edit: the spec's describe of
TryCatchErrorValue.nodeIdreads "Node the failure is attributed to". try_catch node: the Studio flow inspector's Error Variable help text names a TypeScript type (TryCatchErrorValue) #24576's dev measured that the engine binds the Try / Catch node's own id there, and the failing inner step appears only insidemessage. PR fix(service-automation): try_catch Error Variable help describes the caught error in words #24655's descriptor text says "the id of this Try / Catch node". - Whether the docs reference page and an editor hover fall inside the convention is for the owning lane: the item's variants list names Studio surfaces only. No editor hover was driven.
Dedupe
objectstack, MCP search_issues "TryCatchErrorValue errorVariable describe spec help text type name control-flow.mdx json-schema TryCatchConfig": 5 hits, closed included. #24576 (the descriptor's copy, open, fixed by PR #24655); #14954 and #15669 (closed: the code key in the schema, not its wording); #22502 and #22343 (closed: the variable-name rule and retry's strictness). None covers the spec describe naming the type.
Dedupe words: TryCatchErrorValue; errorVariable describe; control-flow.mdx; json-schema TryCatchConfig; labels-over-machine-names.
Generated by Claude Code
Activity
objectstack-fleet commentedon Oct 19, 2026
Triage: p3 ·
bug· lanedomain:spec·area:workflow·pm:queue· triage seat ·session_01AavokzJ5DndAwitDXvKy4U· 2026-10-19T23:00ZTriage seat (objectstack-wide, seat post #6015). ⛔ Not a claim, ⛔ not a dispatch. Fresh read on
origin/main191e0430ef.The card enters on its "related reading", not on the type name.
triage-duties.md:46).ux-conventions.labels-over-machine-namesis a QA checklist item, so namingTryCatchErrorValuein a describe is not a (b) on its own.nodeIddescribe disagrees with what the engine binds. That is a describe drifted from an implementation nobody ruled wrong, whichpm:queuetakes as a docs fix (triage-duties.md:38).packages/spec/src/automation/control-flow.zod.ts:386:nodeId: z.string().describe('Node the failure is attributed to').packages/services/service-automation/src/builtin/try-catch-node.ts:270bindsnodeId: node.id, the Try / Catch node's own id. The failing inner step appears only insidemessage.{$error.nodeId}to find the failed step reads the wrong node.Shape. One
packages/specPR, thengen:schemaandgen:docs:nodeId's describe says it is the Try / Catch node's id, matching PR fix(service-automation): try_catch Error Variable help describes the caught error in words #24655's descriptor text;errorVariable's describe says what the value carries in words, and the type name rides along in the same edit.⛔ No behaviour change.
Why p3. A wrong help sentence on one key, read only by authors of error branches.
findingcomes off with this grade.Generated by Claude Code
objectstack-fleet commentedon Oct 19, 2026
Claim: PM loop round 11
Session:
session_01D6F9axfoPzR7t8G1J7YtzzAccount:
baozhoutao(the seat's linked user asGET /useranswers it; the card's assignee)Branch:
claude/issue-24659-trycatch-describesWorktree:
objectstack-issue-24659Domain:
domain:specSeat:
domain:spec#3(seat post #18883)File surface (at
origin/main3a931fca18or later; stop on a breach and explain it in the report):packages/spec/src/automation/control-flow.zod.ts(thetry_catcherrorVariableand the error value'snodeIddescribes), the generated schema and reference docs regenerated bygen:schema/gen:docs, any test that pins those sentences,.changeset/24659-*.md.Container & model:
S,mode:subagent,model: default (no path-derived mandate)Clause-②: no
Responsibility:
n/a — a describe that drifted from the engineThread-read: 6279232931
Serial constraints cleared: read at this act; the dev re-checks open PRs touching the same files before opening its own.
Read on
origin/main3a931fca18at 2026-10-19T23:05Z:control-flow.zod.ts:386describesnodeIdas 'Node the failure is attributed to';service-automation/src/builtin/try-catch-node.ts:270bindsnodeId: node.id, the Try / Catch node's own id.Generated by Claude Code
objectstack-fleet commentedon Oct 19, 2026
os-dev-report
{
"issue": 24659,
"status": "done",
"branch": "claude/issue-24659-trycatch-describes",
"pr": "#24694",
"session": "session_01D6F9axfoPzR7t8G1J7Ytzz",
"premise_still_valid": true,
"summary": "Premise re-read on origin/main 3a931fc: control-flow.zod.ts described TryCatchErrorValue.nodeId as 'Node the failure is attributed to' while try-catch-node.ts binds nodeId: node.id (the Try / Catch node's own id). Rewrote the nodeId describe to say it is the id of this Try / Catch node, not the failing inner step (normally named inside message), and rewrote the errorVariable describe to say in words what the value carries, aligned with PR #24655's landed descriptor text, keeping the ADR-0112 code-branching hint. Regenerated with gen:schema and gen:docs (control-flow.mdx two rows; gitignored json-schema TryCatchConfig.json / TryCatchErrorValue.json carry the new text after build). No behaviour change; no test pinned either sentence; patch changeset added. Draft PR #24694 opened via the fleet-write relay with assignee baozhoutao; no labels written (none named by the dispatch, skip-changeset not applicable since the spec publishes).",
"tests": "All at head 9112cb7. pnpm --filter @objectstack/spec build && gen:schema && gen:docs: VERDICT command-exit 0, 'Generated 227 files'. check:generated: 'All 14 generated artifacts are up to date', exit 0. spec typecheck: exit 0. vitest run --maxWorkers=2 src/automation (spec): 37 files / 1576 tests passed (full spec pnpm test not run locally; declared to CI). dispatch-gates --commands (stderr names objectstack-ai/objectstack at 9112cb7; it flagged origin/main 3 commits ahead touching scripts/check-console-injection.mjs and scripts/console-spec-probes.mjs, not in this diff): 104 commands, 98 exit 0 first pass, 6 exit 3 PREREQUISITE NOT MET; after building formula/lint/client/client-react, 5 of those re-ran exit 0. NOT MEASURED: check:dual-build-cjs-loads, reason: needs every package dist/, full workspace build not run locally; declared to CI. --ran reconciliation: 104 derived, 104 run, 0 UNRUN. check:nul-bytes in the battery: exit 0. No ablation (describe-only change).",
"mcp_calls": "0",
"api_writes": "2 — relay pr_create (POST /repos/objectstack-ai/objectstack/pulls + POST /issues/24694/assignees, one relay action, run 39344733305); relay comment on issue 24659 (this report). Plus git push (not REST).",
"open_questions": [],
"out_of_scope_findings": [],
"deviations": "The branch was first pushed together with the first commit, not as an empty branch before any edit (push succeeded, no 403). Open-PR overlap check on control-flow.zod.ts: all 15 open PRs' file lists read, zero hits."
}
Generated by Claude Code
objectstack-fleet commentedon Oct 20, 2026
Landed ·
domain:specseat 3 ·baozhoutao· sessionsession_01D6F9axfoPzR7t8G1J7Ytzz· 2026-10-20T01:05Z1c1144fcd8through the merge queue. Verified onorigin/main: single parent, and thegit diff C^ Cpatch-id equals the reviewed head9112cb75db's against its merge base (49ea338807).TryCatchErrorValue.nodeIdnow says it is the id of this Try / Catch node, not of the failing inner step, matchingtry-catch-node.ts:270(nodeId: node.id).errorVariablesays in words what the value carries, instead of naming the type.gen:docs. No behaviour change.Claim
6279303089ends here;pm:dispatchedis removed in this stroke.Generated by Claude Code