GitHub - DanMcInerney/architect-loop: Claude Fable 5 as architect, GPT-5.5 Codex as builder, the repo as memory - a research-backed Claude Code skill for the cross-vendor agent loop · GitHub
/" data-turbo-transient="true" />
Skip to content
Search or jump to...
Search code, repositories, users, issues, pull requests...
-->
Search
Clear
Search syntax tips
Provide feedback
--><br>We read every piece of feedback, and take your input very seriously.
Include my email address so I can be contacted
Cancel
Submit feedback
Saved searches
Use saved searches to filter your results more quickly
-->
Name
Query
To see all available qualifiers, see our documentation.
Cancel
Create saved search
Sign in
/;ref_cta:Sign up;ref_loc:header logged out"}"<br>Sign up
Appearance settings
Resetting focus
You signed in with another tab or window. Reload to refresh your session.<br>You signed out in another tab or window. Reload to refresh your session.<br>You switched accounts on another tab or window. Reload to refresh your session.
Dismiss alert
{{ message }}
DanMcInerney
architect-loop
Public
Notifications<br>You must be signed in to change notification settings
Fork
Star<br>27
main
BranchesTags
Go to file
CodeOpen more actions menu
Folders and files<br>NameNameLast commit message<br>Last commit date<br>Latest commit
History<br>21 Commits<br>21 Commits
assets
assets
skills
skills
tests
tests
.gitignore
.gitignore
DESIGN.md
DESIGN.md
LICENSE
LICENSE
README.md
README.md
install.ps1
install.ps1
install.sh
install.sh
View all files
Repository files navigation
architect-loop
Claude Fable is the architect — it designs every slice, freezes the<br>acceptance gates, and judges the results. GPT-5.5 Codex is the builder and<br>researcher — it does all the engineering and all the web research, in<br>parallel, unattended, for hours. Two Claude Code skills that run this<br>cross-vendor loop on the flat-rate subscriptions you already have — no API<br>keys, no token bills.
Install (30 seconds)
= 0.133)">git clone https://github.com/DanMcInerney/architect-loop<br>cd architect-loop && ./install.sh # Windows: .\install.ps1<br>npm i -g @openai/codex@latest # the builder (Codex CLI >= 0.133)
./install.sh --project installs to the current repo only instead of<br>globally. You need Claude Code on any paid<br>plan and the Codex CLI signed into a ChatGPT plan.
Use (two commands)
# the research loop">/architect # the build loop<br>/architect-research # the research loop
/architect runs one work block: judge the last run, spec the next slice,<br>dispatch builders. /architect-research is for when you're still deciding<br>what to build — its cited report feeds the build loop's PRD.
/architect
One short Fable session per work block — judgment only, it never writes code:
Spec + gates first. Fable specs a one-PR slice, splits it into 1–4<br>lanes with provably disjoint file sets, and commits the acceptance gates to<br>docs/gates/ before any builder starts. Gates are read-only; a builder<br>edit to a gate file fails the slice automatically.
Parallel isolated builders. One fresh codex exec (xhigh) per lane,<br>each in its own git worktree. Builders must argue with the spec before<br>building (silent compliance = defect), build only their declared files,<br>and report raw results — they physically can't commit (the sandbox<br>protects .git).
Fable judges and integrates. It runs the gate commands itself (builder<br>claims are hearsay), reads the diff against the spec's intent (passing<br>tests ≠ mergeable work), then commits and merges passing lanes. Judgment<br>happens in a fresh session — cross-context review measurably beats<br>same-session review.
The repo is the only memory. docs/HANDOFF.md (a short table of<br>contents, pruned every session), docs/gates/, docs/lanes/, git<br>history. Not in the repo = didn't happen.
Supervision built in. Liveness checks on dispatched runs, stall triage<br>(diagnose the child process tree, kill the narrowest thing), explicit<br>timeouts on every long command.
/architect-research
Scout-first, like the production deep-research systems — no fixed lane<br>taxonomy:
A cheap Codex scout maps the topic (~10 searches): canonical<br>terminology, the load-bearing systems and papers, the named people, the<br>topic's natural fault lines. Skipped for comparisons and fact-finds.
Fable designs 3–6 topic-specific lanes from the scout's map, drawing<br>per-source-class tactics from a library (academic citation snowballing,<br>dependents-not-stars repo evidence, emerging-vs-hype gating, production<br>pattern mining, expert tracking) — checked for overlap and gaps before<br>dispatch.
Parallel Codex researchers run under hard budgets: search caps, ≤5<br>subjects per lane, saturation stop, strict findings discipline (URL + date
quote + confidence tag; NOT FOUND beats inference; no recommendations).<br>Expert opinion runs as a second wave, roster-seeded by the first.
Fable verifies and writes. ≥2 independent sources per load-bearing<br>claim, adversarial falsification searches, citations only from...