fix(story-mode): what a non-technical reader would believe, from the cold-reader review

Review round 5 had an agent read only the comic of the nine real sessions
(20 turns), write down what it believed happened, and only then compare
with the transcripts. Three panels told it something false, one hid the
failure the user asked to see:

- A push that went through, followed by a check that failed in the same
  command, drew as a failed push; the output (`[new branch]`, `->`) now
  decides: "Claude saves the work in 2 commits and uploads it, but a step
  after that fails."
- A formatter's "would be reformatted" preview, printed before it fixed
  the file, made a clean run a failure: the last verdict wins.
- A blocked command ended in confetti and "Mission complete!": a run panel
  where everything failed says "Claude tries a command, and it is
  blocked.", and a turn where nothing worked ends "Turn over" / "Nothing
  Claude tried worked, so the job is not done." with a neutral robot.
- The loop stamped only the final verdict (POW ×2), so a first run that
  failed on purpose never showed. It now marks every test run red or green,
  and says it: "Fix and test: the first run fails, the second passes." /
  "Build and test, 4 rounds. A check fails once, then everything passes."

Also:
- Files written inside a check or a save reach the plain caption ("Claude
  writes 4 files; the tests pass.").
- Balloons keep the next sentence when the first reports a problem ("…left a
  stray file. I'll remove it…") or is very short ("Not done. The harness
  blocked sleep 65…"); Claude's last words keep a caveat that follows them
  ("Nothing is committed because this folder isn't a git repository.").
- The advisor never shares a panel with helpers that do work; ToolSearch
  reads "Claude gets the tools it needs ready."; Codeman's own case file is
  not drawn as a file Claude read; looking around shows what Claude meant
  (the call's description), not raw commands; a montage at the start of a
  turn reads "Claude gets started".
- Words: boards count tests ("19 tests"), "uploads it" replaces "sends it
  up", and the prompt balloon keeps 400 characters.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
This commit is contained in:
Codeman maintainer committed 2026-10-18 12:32:40 +02:00
1 parent 0effbb9ce4
commit 814bec739d
14 files changed
+307 -109

No files matched your search

@@ -17,14 +17,14 @@
"ok": null,
"story": "You give Claude a job.",
"director": "Prompt · 52 words",
"balloon": "Build a small Python command-line tool wcount.py in this folder that prints the line, word and character counts of the files given as arguments, like wc, with a --json flag that prints the counts as JSON, plus pytest te…",
"balloon": "Build a small Python command-line tool wcount.py in this folder that prints the line, word and character counts of the files given as arguments, like wc, with a --json flag that prints the counts as JSON, plus pytest tests in test_wcount.py; run pytest and fix anything that fails until all tests pass.",
"tests": null,
"events": 1
},
{
"kind": "investigate",
"ok": null,
"story": "Claude reads 1 file to find its way around.",
"story": "Claude looks around the project.",
"director": "Bash ×2 · ls -la && git log --oneline -5 2>&1; git branch -a; python3… · 1 error",
"balloon": null,
"tests": null,
@@ -42,7 +42,7 @@
{
"kind": "loop",
"ok": true,
"story": "Fix and test, 2 rounds. It passes in the end.",
"story": "Fix and test: the first run fails, the second passes.",
"director": "3 checks · 5 edits · +218 −1 · last: ok",
"balloon": "All 19 tests pass, ruff check is clean, and the output matches real wc column for column.",
"tests": {
@@ -54,7 +54,7 @@
{
"kind": "save",
"ok": true,
"story": "Claude saves the work.",
"story": "Claude writes 1 file and saves the work.",
"director": "git commit -q -m \"chore: ignore Python and pytest caches · writes 1 file",
"balloon": null,
"tests": null,