WORK_CYCLE.md covered only the end of the cycle. A cycle has two ends, and
the same six questions kept having to be asked out loud on every piece of
work. They are now mandatory in every plan, each justified by something this
household has actually paid for:
unified code eight copies of secrets.sh once existed here and five of
seven could not detect the most common secret shape --
INCLUDING THE TEMPLATE, so every project scaffolded from
it inherited a blind scanner
error handling the recurring fault is the silent pass, not the crash
logging an append-only log is unbounded by construction
blind spots named ones get fixed
landmine fixes the trap found while passing is cheapest to fix while
passing
hardcode little derive it, or justify the constant
GUARDS.md was five sections behind the project that has been learning; 11,
12 and 13 are backported, genericised to match the template's style. 13 is
the narrow form of the sixth rule, and WORK_CYCLE now cites it -- so
backporting it is what makes that citation true rather than a broken
reference.
Also adds the three files an agent reads BEFORE it reads docs/:
CLAUDE.md short, and it POINTS at DOC_TRUST_MAP rather than
repeating it -- a second copy of the map is the
failure that map exists to prevent. Carries the
exit-code table, the commit gates, and two standing
instructions: flag what looks wrong even when it is
not what you were asked about, and never print a
credential -- not from a file, not from a command's
output, not from a config subtree "with the secrets
filtered out", because that filter has failed before
by matching key NAMES while the secret sat inside an
object whose name was innocent
.claudeignore excludes artifacts and NEVER docs/. The Command
Center reads this repository's documents at a commit;
a generic ignore file that sweeps "documentation" or
"data" starves both the agent and the reconcile, and
everything still runs, just blind
.claude/settings.json deny rules in the double-slash absolute form. A
tilde-style rule looks right in review and silently
matches nothing. It closes the Read TOOL only -- a
shell reads a file a hundred ways -- so it catches
the accidental read, not the determined one
scaffold.sh gains a ROOT array for the three, kept apart from DOCS so the
H1-plus-status-block check stays meaningful rather than being loosened into
a warning that is always wrong (GUARDS.md 5).
Verified: scaffold --dry-run into a scratch repo creates 22 files including
all three, with no HEADERLESS warning; doc-claims passes with 124 claimed
paths, all present.
Does not touch docs/architecture/scripts/secrets.sh, which carries someone
else's uncommitted improvement.
Two kinds of project arrive at this template and neither had a document written
for it. A new repository needs a first act. An existing one -- which is most of
them, since eight were surveyed and one had adopted anything -- needs a merge,
with documents and conventions already in place that must survive it.
What existed was README's "Adopting it": seven steps aimed at a person reading
top to bottom, silent about existing projects, and opening with the one form
that destroys work:
cp -r Projects/Template/docs <your-project>/docs
cp -r overwrites. scaffold.sh exists precisely because that is unsafe, and its
header argues the case at length: a clobbered document is work nothing notices
is gone, silent when it happens and silent afterwards. An existing project
following step 1 literally replaced its own architecture notes with
placeholders. The instruction and the tool disagreed and the instruction was the
dangerous one.
Both documents hold one prompt each, written to be handed to an agent whole, and
both open by putting it in PLAN MODE with nothing changed until the plan is
approved. Each plan must name what will be created, what kept, what deleted,
which scripts the project adopts -- and what will be left undone, by name. That
last is the one that gets skipped: an adoption quietly missing the webhook or
half the status headers looks identical to a complete one from outside.
The existing-project prompt is the harder half. Its spine is survey, reconcile,
destroy nothing: the survey IS the plan phase and is entirely read-only; the
scaffold's "kept" list is the merge worklist; existing prose is the project's own
knowledge and the template supplies shape, not content; and any markdown backlog
moves into the tracker and is deleted in the same commit, because two records of
what is open is the failure the whole convention exists to prevent.
Verified rather than asserted, since the whole document rests on it: a scratch
project with three written documents, scaffolded into, reported 16 created and 3
kept, and all three pre-existing files were byte-identical afterwards with their
prose intact.
README's section becomes a pointer to the two, since three descriptions of
adoption in one repository is the two-records failure with an extra copy, and the
sentence about the Command Center's generated prompt now says which is the
source rather than leaving a reader to pick.
Also corrected: README's `Governs: Projects/Template/**` was true when this was a
folder inside Projects/ and matched nothing once it became a repository root.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
null