Three defects in drive-pr's outcome wording, all pre-existing, surfaced by a whole-unit read of "Disposing of Every Finding" and the frontmatter description.
Outcome 2's scope is wider here than on the two surfaces that state it. drive-pr says "Not real, or real but out of scope here, so decline in the thread with evidence", where pr-review-conduct and local-strict-review both say "real but structurally out of scope" and illustrate it with ownership sitting elsewhere. "Out of scope here" reads naturally as out of this pull request's scope, which is outcome 4's territory ("real and worth doing later, so file the issue first"). The looser wording licenses closing with a decline a finding the other two surfaces require be deferred behind a filed issue, and the decline is the direction where nothing survives the merge.
Outcome 3 carries a trigger its owner does not list, and drops the ordering rule. "or the agent genuinely does not know which of the above applies" is not among pr-review-conduct's four escalation triggers (design trade-off, recurring finding, judged real but not to be fixed, architectural redesign), and the bullet omits that outcome's precondition, "Reach for this only once outcome 2 is ruled out", which local-strict-review does carry.
The frontmatter enumerates four of the five outcomes as if complete. "fix it, decline it with evidence, defer it behind a filed issue, or put the call to the maintainer" omits outcome 5 (a recurring finding, so fix the class), while the body calls them "pr-review-conduct's five outcomes".
Also in the frontmatter: "Ends at develop merged, or at a promotion PR meeting the pr-review-conduct Merge Gate." Merge Gate item 5 is "The maintainer has given explicit permission to merge", which this skill's own "What Invoking This Skill Authorizes" says it never obtains for the promotion PR. The body's real exit is steps 8 and 9: no open finding, checks green on the current head, "Do not merge it." As written the description names an end state the drive cannot reach in its own seat.
And in the body: "Real, so fix it, then step 2's own order again before replying with the fixing commit SHA (outcome 1)" leaves "step 2" unresolvable from inside the unit, where local-strict-review qualifies the same reference as "drive-pr's Drive Loop step 2".
Raised by the local-strict-review carried-content passes on the pull request that added backlog-burndown, and deferred there under that skill's own review-round budget.
Three defects in
drive-pr's outcome wording, all pre-existing, surfaced by a whole-unit read of "Disposing of Every Finding" and the frontmatter description.Outcome 2's scope is wider here than on the two surfaces that state it.
drive-prsays "Not real, or real but out of scope here, so decline in the thread with evidence", wherepr-review-conductandlocal-strict-reviewboth say "real but structurally out of scope" and illustrate it with ownership sitting elsewhere. "Out of scope here" reads naturally as out of this pull request's scope, which is outcome 4's territory ("real and worth doing later, so file the issue first"). The looser wording licenses closing with a decline a finding the other two surfaces require be deferred behind a filed issue, and the decline is the direction where nothing survives the merge.Outcome 3 carries a trigger its owner does not list, and drops the ordering rule. "or the agent genuinely does not know which of the above applies" is not among
pr-review-conduct's four escalation triggers (design trade-off, recurring finding, judged real but not to be fixed, architectural redesign), and the bullet omits that outcome's precondition, "Reach for this only once outcome 2 is ruled out", whichlocal-strict-reviewdoes carry.The frontmatter enumerates four of the five outcomes as if complete. "fix it, decline it with evidence, defer it behind a filed issue, or put the call to the maintainer" omits outcome 5 (a recurring finding, so fix the class), while the body calls them "pr-review-conduct's five outcomes".
Also in the frontmatter: "Ends at develop merged, or at a promotion PR meeting the pr-review-conduct Merge Gate." Merge Gate item 5 is "The maintainer has given explicit permission to merge", which this skill's own "What Invoking This Skill Authorizes" says it never obtains for the promotion PR. The body's real exit is steps 8 and 9: no open finding, checks green on the current head, "Do not merge it." As written the description names an end state the drive cannot reach in its own seat.
And in the body: "Real, so fix it, then step 2's own order again before replying with the fixing commit SHA (outcome 1)" leaves "step 2" unresolvable from inside the unit, where
local-strict-reviewqualifies the same reference as "drive-pr's Drive Loop step 2".Raised by the local-strict-review carried-content passes on the pull request that added
backlog-burndown, and deferred there under that skill's own review-round budget.