Commit 9d755ba
skills(automation): optimization flight — the authored control-flow trio taught and the BPMN trio demoted, notify given its fence, the eval deduplicated, hooks and ETL pointed away (net −1,278 tokens) (#14467)
* skills(automation): deduplicate the send-back eval (AUTO-H-01)
The eval restated SKILL.md's "Send-back for revision (ADR-0044)" section rather
than testing beyond it: the canonical shape, the four lint findings, the
wait-vs-approval_revise rationale and the showcase pointer all had a second copy
here. 2,204 tokens -- 15% of the authored package -- went to a surface with one
real authoring in the whole corpus, while `notify` (20 authorings) got 15.
Delete the Expected-Output fence and the Common-Mistakes rows that restate the
skill; keep Scenario + Validation Criteria, which is what a grader needs. The one
row that did NOT restate SKILL.md (a "revise mode" re-suspend of the approval
node) survives as criterion 7, so no graded fact is lost.
1,329 -> 550 tokens (ceiling 1,329).
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LraLgQVGq8egUwfYZpbYt1
* skills(automation): teach the authored control-flow trio, demote the BPMN one
The package taught only `parallel_gateway` / `join_gateway` / `boundary_event`
-- 0 authorings anywhere in the corpus -- and never mentioned the ADR-0031
structured constructs the corpus actually uses (`loop` 3, `try_catch` 2,
`parallel` 2). `notify`, the most-used node type at 20 authorings, had one table
cell and no config keys, including no mention of its documented silent no-op.
ADD (funded, each claim verified at source):
AUTO-F-01 `parallel` / `try_catch` / `loop`-container rows; the gateway trio
marked BPMN-interop lowering, not author-facing.
AUTO-F-02 capability -> surface table, adding `approvals` and `messaging`.
AUTO-E-04 `notify` config fence + the messaging-absent silent success.
ported flow-filter precedence + the static-checkability split (this
package is the anchor; the query flight deleted its copy).
AUTO-G-02 `autolaunched` + a `record-*` binding behaves like `record_change`.
AUTO-E-01 os:check marker on the schedule-flow example (265 -> 266 blocks).
AUTO-A-02/A-03 `jobs` / `webhooks` collections and screen-flow routing in the
description.
DELETE / MERGE: AUTO-A-01 (ETL, retired at spec 17), AUTO-B-03, AUTO-B-06,
AUTO-B-07, AUTO-C-01, AUTO-C-02, AUTO-C-03, AUTO-C-04, AUTO-D-01..D-09,
AUTO-E-02. REORDER (free): AUTO-B-01, AUTO-B-02.
12,768 -> 12,269 tokens (ceiling 12,768; every addition paid in this file).
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LraLgQVGq8egUwfYZpbYt1
* skills: regenerate the skill index after the automation description edit
Generator output only -- `pnpm --filter @objectstack/spec gen:skill-docs`.
`check:skill-docs` is a drift gate, so the AUTO-A-01/A-02/A-03 description edit
forces both files.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LraLgQVGq8egUwfYZpbYt1
---------
Co-authored-by: Claude <noreply@anthropic.com>1 parent 310dd01 commit 9d755ba
4 files changed
Lines changed: 281 additions & 377 deletions
File tree
- content/docs/ai
- skills
- objectstack-automation
- evals/approvals
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
47 | 47 | | |
48 | 48 | | |
49 | 49 | | |
50 | | - | |
| 50 | + | |
51 | 51 | | |
52 | 52 | | |
53 | 53 | | |
| |||
117 | 117 | | |
118 | 118 | | |
119 | 119 | | |
120 | | - | |
| 120 | + | |
121 | 121 | | |
122 | | - | |
| 122 | + | |
123 | 123 | | |
124 | 124 | | |
125 | 125 | | |
| |||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
33 | 33 | | |
34 | 34 | | |
35 | 35 | | |
36 | | - | |
| 36 | + | |
37 | 37 | | |
38 | 38 | | |
39 | 39 | | |
| |||
0 commit comments