Content
67%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is a well-structured, highly actionable reference with an unusually disciplined mutation workflow (single reviewed transaction, approval-hash gating, honest limits on unverified behavior). Its main weaknesses are token overhead from version/date provenance detail and a digressive date-arithmetic evidence narrative, plus the absence of explicit error-recovery loops in the workflow.
Suggestions
Move the documentation-provenance narrative in 'Date arithmetic needs native evidence' (changelog vs. help-docs comparison, API commit hash) into a referenced notes file, keeping only the candidate expressions and the instruction to test them via native-conformance.md.
Consolidate the version minimums and check dates now spread across 'Version boundary' and 'Date arithmetic' into one clearly labeled version-compatibility section so time-sensitive detail is centralized and cheap to skim.
Add explicit error-recovery branches to the workflow, e.g. what to do when transaction inspect reports validation errors (fix the bundle, re-inspect, only then re-apply) so the feedback loop is stated rather than implied.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly efficient (terse rule bullets, compact schema examples) but carries time-sensitive detail outside any deprecated section ("Checked 2026-10-17", versions 1.9+/1.10+/1.14+, a pinned commit hash) and a ~25-line date-arithmetic section of documentation-provenance narrative ("The recorded excerpts of the API types at the reviewed commit do not include a Duration definition...") that could be tightened. This is more than the minor trimming of the level-4 anchor, so it sits at mostly-efficient-with-some-unnecessary-explanation. | 3 / 5 |
Actionability | Guidance is mostly executable: complete paste-ready YAML for filters, formulas, table/cards/list/kanban views, and embeds, plus concrete transaction inspect/apply commands. It falls short of fully copy-paste ready because PRODUCT_ROOT is a placeholder, $BUNDLE/$VAULT construction is delegated to an external file, and the date formulas are explicitly "candidates for an isolated test, not a compatibility guarantee". | 4 / 5 |
Workflow Clarity | The 8-step workflow has a clear sequence and strong explicit checkpoints (validate YAML/quoting/properties, preview before mutation, inspect-then-apply gated on an approval SHA256, and separating static validation from native rendering). It lacks the explicit error-recovery branches ("if errors: fix and re-validate") that distinguish the top anchor, so it fits clear-sequence-with-most-checkpoints. | 4 / 5 |
Progressive Disclosure | Section structure is good and external references are clearly signaled with explicit resolution rules ([operation-transactions.md](../wiki/references/operation-transactions.md), [native-conformance.md](../../docs/native-conformance.md)), with no nested reference chains. It is not top-level because all content lives inline in one ~200-line file with no reference bundle of its own, and the date-arithmetic evidence detail would fit better in a separate reference file. | 4 / 5 |
Total | 15 / 20 Passed |