feat(skills): drive the right side of the V + check tool currency before filing gaps - #156
Merged
Merged
Conversation
…ore filing gaps
Two changes, both from mining the ten unactioned June field reports, and both
TESTED against a real rivet project before being written down.
feature-loop step 3 — close the V when the test lands:
Ten reports converged on one gap: the left side of the V grows fast because
authoring it is natural, while the right side is skipped silently every feature
(one repo: 257 of 269 sw-reqs unverified; another: 20 artifacts, one verifies
link). Five of them independently proposed the same skill under five names.
The cheap path already exists in rivet and no skill mentioned it: a source
marker instead of a hand-authored YAML artifact per test. Verified end-to-end on
a scratch dev-preset project with the INSTALLED binary:
// rivet: verifies REQ-001 in a test
rivet coverage --tests -> REQ-001 1 test marker ./src/lib.rs:6 (verifies)
1/1 requirements have test coverage (100.0%)
rivet verify REQ-001 -> evidence: 1 source verifies marker(s); status
changed on disk to `verified`
rivet verify REQ-002 -> REFUSES: "no verifying evidence. Add an incoming
verifies link ... or a marker, then re-run"
That refusal is what makes this a mechanical check rather than a prose step,
which was #94's explicit constraint on any new guidance ("don't add prose-only
skills — they get skipped under pressure").
report-tool-friction 2b — version currency before a capability gap:
A missing capability is the one friction class that expires. Three measured
instances of the waste: rivet#748 filed 13 days AFTER the type it asks for
shipped (reporter on dev@0.2.0, fixed in 0.3.0 on 2026-07-17); a session pinned
to the oldest plugin version, hiding eight skills; this machine running rivet
0.28.0 against a v0.32.0 release. Bugs and crashes still get filed immediately —
only missing-feature claims need the thirty-second version check.
feature-loop 0.1.0 -> 0.2.0; report-tool-friction 0.1.0 -> 0.2.0; plugin 0.18.0.
Refs #84, #88, #89, #90, #91, #92, #93, #94, #95
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
feat(skills): drive the right side of the V + check tool currency before filing gaps
Two changes, both from mining the ten unactioned June field reports, and both
TESTED against a real rivet project before being written down.
feature-loop step 3 — close the V when the test lands:
Ten reports converged on one gap: the left side of the V grows fast because
authoring it is natural, while the right side is skipped silently every feature
(one repo: 257 of 269 sw-reqs unverified; another: 20 artifacts, one verifies
link). Five of them independently proposed the same skill under five names.
The cheap path already exists in rivet and no skill mentioned it: a source
marker instead of a hand-authored YAML artifact per test. Verified end-to-end on
a scratch dev-preset project with the INSTALLED binary:
// rivet: verifies REQ-001 in a test
rivet coverage --tests -> REQ-001 1 test marker ./src/lib.rs:6 (verifies)
1/1 requirements have test coverage (100.0%)
rivet verify REQ-001 -> evidence: 1 source verifies marker(s); status
changed on disk to
verifiedrivet verify REQ-002 -> REFUSES: "no verifying evidence. Add an incoming
verifies link ... or a marker, then re-run"
That refusal is what makes this a mechanical check rather than a prose step,
which was #94's explicit constraint on any new guidance ("don't add prose-only
skills — they get skipped under pressure").
report-tool-friction 2b — version currency before a capability gap:
A missing capability is the one friction class that expires. Three measured
instances of the waste: rivet#748 filed 13 days AFTER the type it asks for
shipped (reporter on dev@0.2.0, fixed in 0.3.0 on 2026-07-17); a session pinned
to the oldest plugin version, hiding eight skills; this machine running rivet
0.28.0 against a v0.32.0 release. Bugs and crashes still get filed immediately —
only missing-feature claims need the thirty-second version check.
feature-loop 0.1.0 -> 0.2.0; report-tool-friction 0.1.0 -> 0.2.0; plugin 0.18.0.
Refs #84, #88, #89, #90, #91, #92, #93, #94, #95
Co-Authored-By: Claude Opus 4.8 (1M context) noreply@anthropic.com