improvise-28ri, from the #211 review at 6ccad4b: the recipe's grep has no
-r, so the skills' symlinks arrive as directory operands and are never
read. -d skip only silences "Is a directory", and sort -u would absorb a
duplicate anyway. Measured with BSD grep 2.6.0 and GNU grep 3.12: 3092
matches either way, 462 distinct ids. GNU grep -r follows those
symlinks (3107), so -r must not be added.
improvise-oga4: "Shipping a wave"'s opening and "Repository policy over
the Beads blocks" now point at the ADR exception. The wave SKILL.md
pointer is unwrapped, so a grep for "ADR pull requests" finds it.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
The user's ruling: "An ADR isn't approved until I say it is." CI, the
reviews and the adversarial review stay required as gates, but none of
them is approval. SKILL.md's "not on a review" pointer is reworded to
match.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
improvise-28ri: -d skip only silences "Is a directory" on the skills'
symlinks, which git ls-files lists as files. grep -r never follows them,
so nothing was ever read twice. The "spells no id" sentence now says what
the recipe actually cites.
improvise-oga4: the review-reply and head-SHA rules live only in
agent-rules "GitHub" (plus the after-merge clause moved there); AGENTS.md
points to it, and the code-review skill names it. An ADR's "Approved" is
an extra gate on top of CI, every review answered and the adversarial
review.
improvise-diho: agent-rules rule 7, pin gh/main to a SHA at task start.
Also: build a test-name list in a fresh CARGO_TARGET_DIR, because
git archive's commit mtimes make a reused one return a stale list.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
- agent-rules "GitHub": authors reply on the PR to every finding they
address (commit, or why declined); reviewers read existing reviews first
with scripts/pr-reviews, answer every Codex review on its thread, and
state the head SHA they reviewed. SKILL.md points reviewers there.
- AGENTS.md "ADR pull requests": numbered open questions, the user answers
on the PR, "What the user decided", merge only on "Approved" (PR #122).
SKILL.md "Shepherding" points to it.
- Fix a pointer to the renamed policy section that wrapped across a line
break and so escaped the rename.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
- User ruling: work is not done until it is in a PR and that PR has
merged. Stated once, in AGENTS.md's repo-owned section outside bd's
markers (renamed "Repository policy over the Beads blocks"); agent-rules
"Do not merge" and SKILL.md "Shepherding" point to it.
- "Who owns what" no longer says kb facts are batched two sentences
before saying they are not: repo-map lines are batched, kb facts are
written as they arrive.
- improvise-8haj: the ID audit takes its file list from git ls-files
(tracked + untracked, minus ignored), so build output such as
harness/serve and other agents' worktrees are not scanned; grep -d skip
passes over the .claude/skills symlinks.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
21x21 grid, 8x8 window, start (10,10): moves land at 10+d, page-scroll at
10+page_rows(8)*d, jumps at 0/20. grid_move's page rule is page_rows(),
which the test reads rather than restating. Doubled move-selection, a
one-row scroll-rows and a page-size-blind page-scroll each fail.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
ApplyAndClearDrill walked pending_edits, a HashMap, and applied each edit on
its own. A rename moves the row's cells, so a measure edit replayed after it
wrote the old key. That left a ghost record there and a stale value in the moved
record. A second rename on the same row always found nothing and was dropped.
Across rows, a swap lost one record's data in either order.
RecordRewrite::of builds each row's final record once. It reads unedited values
off the model before anything is written, and applies every rename to every
cell. RecordRewrite::merged puts every clear before every write, in snapshot row
order. The plain records path calls the same function with its one edit, so
rename_record_coordinate and the drill's own copy are gone (gotcha 19).
Reasoning: context/code/command.md, RecordRewrite.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
The evidence file is gh/main plus this branch's five entries, inserted at
their original anchor. command.md no longer claims a client's history
cannot hold a dead name: fresh_view seeds a phantom Default frame
(improvise-lxg0, not fixed here).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
improvise-6yvg reproduced. A JSON import in improvise-command and a legacy
JSON .improv load in improvise-io stored 0.24815714285714285 as
0.24815714285714283 in each crate's own test build. The root binary's
`improvise import` used the same lossy parser. Now float_roundtrip is on every
normal serde_json dependency in the workspace: root, improvise-command,
improvise-io, and the optional test-support dependency of improvise-protocol.
improvise-dz8c: the hand-listed cargo-tree check over three frame-parsing
crates is replaced by tests/json_numbers.rs. It reads
`cargo metadata --no-deps` and names every package, including target-specific
entries, whose normal serde_json dependency lacks the feature. The rule and why
it beats "parses a frame", the build graph, and enabling the feature once in
the protocol crate are in context/code/protocol.md.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
- improvise-rzux: AGENTS.md now states the per-agent PR workflow the wave
skill uses (adversarial review at the head SHA, merge on GitHub, local main
fast-forwards to gh/main) and points to SKILL.md "Shepherding". The
obsolete one-PR-per-wave and local-integration text is cut.
- improvise-x1uk: the Beads integration block is regenerated whole by bd
(bd init, bd setup factory|opencode|mux). It is replaced with bd 1.3.0's
own minimal render, which no longer says git merge origin/main, and the
repository's policy (no team-maintainer opt-in, the TODO exception that
had been hand-edited inside the markers) moves to a section outside them.
- improvise-anmm: the ID audit scans .agents and the tracked .claude files,
excludes the kb skill's example crate name, and anchors its matcher so a
ticket body saying "not found" is not reported as a dead id.
- improvise-xolb: agent-rules' gate list is the test/js/kb CI jobs step for
step, with the conditional jobs (wasm incl. improvise-gp, harness-lint,
build+browser incl. the bundle node tests) and the excluded one named.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
The worker transport wrote "in-page worker · <doc>" when it spawned the
worker, before anything had answered, so a tab whose worker could not start
showed a success-looking status over a blank page. It now writes "starting
the worker · <doc>…", and createShell's arm() takes an `answer` that sawFrame
writes on the first frame, but only while the watchdog is still armed. A
ReadOnly or SeedMissing notice lands BEFORE the Snapshot, and fail() has
already disarmed, so the claim cannot overwrite the precise reason.
The two-tab spec now logs every #status and denies both the premature claim
and "did not answer" before the page clock is walked. A new spec pins the
exact status sequence on the writable path. Reasoning:
context/code/clients.md, "The watchdog".
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
KeyRow::looked_up is private to key_row, whose keys() is the lookup
pages() hands the builders; the reviewer's direct call is E0624. A new
test asserts the drawn keys rows equal help_bindings(), which catches
the remaining self::keys bypass.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
scroll-rows and page-scroll derive their landing from their argument as
move-selection does; any bound command that moves the cursor with no
stated landing fails; the numeric floor is gone. Reversed Ctrl-d/u and
PgUp/PgDn, and a dropped jump table with $ sent to column 0, both fail.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VdfEGXA7aJ6ftMsCc9PJze
Edward Langley