# Claude Code First changelog — Alegauss

Every entry in Claude Code First's roadkeep ledger, dated by the commit that wrote it.

- 2026-10-20 · CCF80: Rule pages print literal asterisks around every normative sentence that holds an italic term
- 2026-10-20 · CCF79: The site header links neither alegauss.github.io nor the GitHub repository, unlike every other alegauss site
- 2026-10-20 · CCF78: markdownlint lints the governed docs/DEFERRED.md, and fails on a blank line roadkeep wrote there
- 2026-10-20 · CCF77: The guide and specification pages carry their own 892-line stylesheet, outside the design system
- 2026-10-20 · CCF76: The site shows no mark of its own: no favicon, no logo in the header, no social card
- 2026-09-25 · CCF75: the files-over-APIs open question names what would settle it, and nothing has been measured against it
- 2026-09-25 · CCF74: the method admits no evidence of what a project's own sessions cost, so the session-budget question cannot be settled
- 2026-09-25 · CCF73: the guide's lessons say nothing about what an agent-facing surface costs, and the page types its lesson count
- 2026-09-25 · CCF72: no rule asks a product to benchmark canonical tasks through its agent surface against the generic path
- 2026-09-25 · CCF71: the files-over-APIs and the per-project session token budget claims are recorded nowhere as open questions
- 2026-09-25 · CCF70: no rule says when a capability belongs on a CLI verb rather than on an MCP tool list
- 2026-09-25 · CCF69: no finding records putting capabilities on a CLI rather than an MCP tool list where the consuming agent has a shell
- 2026-09-25 · CCF68: no finding records that a task-shaped agent path cost a fraction of the generic API's tokens, or where the gap came from
- 2026-09-25 · CCF67: the site opens on a table of rules, and a reader setting up a project has no plain-language way in
- 2026-09-25 · CCF66: CI still excuses winwright from the pointer check after its pin was pushed
- 2026-09-25 · CCF65: this repository declares no line terminators in .gitattributes, so it fails its own rule EP-2
- 2026-09-25 · CCF64: the citation resolver takes about five minutes over 795 pointers, so it will be the gate a commit skips first
- 2026-09-25 · CCF63: the resolver cannot check a quote that holds inline code or runs across comment-prefixed source lines
- 2026-09-25 · CCF62: the repository has no README telling a newcomer what the spec is, how to read it and how to adopt it
- 2026-09-25 · CCF61: the work has no license, no citation metadata and no statement of how it was authored
- 2026-09-25 · CCF60: an agent in another repository cannot consume the spec cheaply: no llms.txt and no page per rule
- 2026-09-25 · CCF59: no process admits new evidence or a new rule, so the spec would either freeze or grow by opinion
- 2026-09-25 · CCF58: the specification has no version, so an audit cannot say which text it judged a project against
- 2026-09-25 · CCF55: shio has never been audited against the specification, so its conformance is asserted, not measured
- 2026-09-25 · CCF54: winwright has never been audited against the specification, so its conformance is asserted, not measured
- 2026-09-25 · CCF53: freewilly has never been audited against the specification, so its conformance is asserted, not measured
- 2026-09-25 · CCF52: polyweave has never been audited against the specification, so its conformance is asserted, not measured
- 2026-09-25 · CCF51: roadkeep has never been audited against the specification, so its conformance is asserted, not measured
- 2026-09-25 · CCF50: a project already Claude Code first that drifted has no path back except rereading the whole spec
- 2026-09-25 · CCF49: an existing codebase adopting the practice has no procedure, and Shio's transition lives only in its history
- 2026-09-25 · CCF48: starting a project Claude Code first has no ordered procedure, so the order the five learned in is repeated
- 2026-09-25 · CCF47: a new project has no starting kit, so each of the five wrote its own settings, hooks and skills
- 2026-09-25 · CCF46: a deliberate deviation from a rule cannot be recorded, so it reads as drift in every audit
- 2026-09-25 · CCF45: an audit report has no fixed format, so two audits of one project cannot be compared over time
- 2026-09-25 · CCF44: the rules that need judgement cannot be audited without an agent reading the whole spec
- 2026-09-25 · CCF43: automatable rules have no checker, so each audit rederives by reading what a script could measure
- 2026-09-25 · CCF42: traceability is unchecked: a rule citing no finding, or a finding citing no evidence, would pass
- 2026-09-25 · CCF41: rules exist only as prose, so no tool can list them, filter them by level or check a repository against them
- 2026-09-25 · CCF40: conformance has no levels, so a project cannot say how far it adopts the spec or what to do next
- 2026-09-25 · CCF39: the practices that resolved those failures are described only inside the projects that invented them
- 2026-09-25 · CCF38: the recurring failures have no names, so a reviewer cannot say which one a project is committing
- 2026-09-25 · CCF37: no form fixes how a pattern or anti-pattern is written, so entries would mix problem, remedy and evidence
- 2026-09-25 · CCF36: no normative chapter covers the prose an agent writes, though model-written prose was measurably detectable
- 2026-09-25 · CCF35: no normative chapter covers several agent sessions working in one repository at once
- 2026-09-25 · CCF34: no normative chapter says what the person decides and what an agent must never certify for itself
- 2026-09-25 · CCF33: no normative chapter covers projects that ship tools, skills or plugins for other agents to use
- 2026-09-25 · CCF32: no normative chapter covers where a session runs: web versus local, shells, encodings and line endings
- 2026-09-25 · CCF31: no normative chapter states a permission and safety posture, and the corpus runs with prompts bypassed
- 2026-09-25 · CCF30: no normative chapter says which rules must be enforced by hooks rather than by instructions
- 2026-09-25 · CCF29: no normative chapter says what an agent must run, keep and report before calling work done
- 2026-09-25 · CCF28: no normative chapter says how an agent turns finished work into commits
- 2026-09-25 · CCF27: no normative chapter says how an agent plans, records and closes work
- 2026-09-25 · CCF26: no normative chapter says what may load on every turn and what must be trigger-loaded instead
- 2026-09-25 · CCF25: terms such as every-turn file, governed file, gate, guard, ledger and verdict are used without definitions
- 2026-09-25 · CCF24: the specification has no frame: no chapter order, no rule addresses and no conformance keywords
- 2026-09-25 · CCF23: no cross-project metrics exist, so claims about scale, cadence and context cost are adjectives
- 2026-09-25 · CCF22: the drift already present in the corpus is uncatalogued, though each case shows prose rules decaying
- 2026-09-25 · CCF21: the projects disagree on attribution, instruction files, permissions and CI scope, and nothing settles it
- 2026-09-25 · CCF20: lessons are scattered over five ledgers, so one is counted five times and a contradiction not at all
- 2026-09-25 · CCF19: no case study describes shio, the only project that adopted the practice after years of history
- 2026-09-25 · CCF18: no case study describes winwright, whose agent surface is shipped for other repositories to use
- 2026-09-25 · CCF17: no case study describes freewilly, the project that works without any instruction file
- 2026-09-25 · CCF16: no case study describes polyweave, the one project governed from its very first commit
- 2026-09-25 · CCF15: no case study describes roadkeep: its timeline, agent surfaces, gates and the incidents that shaped them
- 2026-09-24 · CCF14: no related work is surveyed, so the spec cannot say what is new and what restates ADRs, RFC 2119 or pattern languages
- 2026-09-24 · CCF13: the Claude Code behaviour the spec relies on is cited from memory rather than from versioned documentation
- 2026-09-24 · CCF12: threats to validity are unstated: one author, one harness, one model family, five projects over six months
- 2026-09-24 · CCF11: the field notes were extracted by agents and not one of their claims has been checked against the pinned sources
- 2026-09-24 · CCF10: claims carry no strength, so one incident in one project reads like a pattern seen in all five
- 2026-09-24 · CCF9: an evidence pointer has no grammar, so nothing can check that the path, line or commit it names exists
- 2026-09-24 · CCF8: the corpus is named by path only, so a citation breaks as soon as a source repository moves on
- 2026-09-24 · CCF7: no protocol says how evidence is gathered or judged, so a finding here cannot be reproduced or challenged
- 2026-09-24 · CCF6: a Write over an existing file is accepted silently here, the path that destroyed about 700 lines in pportal
- 2026-09-24 · CCF5: the every-turn files carry no size budget, so they can grow the way Shio's agents.md reached 186 KB
- 2026-09-24 · CCF4: nothing states how work is committed here, so batching, stray files and mislabelled docs commits are all possible
- 2026-09-24 · CCF3: no gate runs on push, so a lint failure or a broken link in the specification reaches main unremarked
- 2026-09-24 · CCF2: the governed docs have no guard here, so a hand edit of ROADMAP.md or IMPROVEMENTS.md passes unnoticed
- 2026-09-24 · CCF1: a session opening this repository is told nothing: no every-turn file states purpose, language or write paths
- undated · CCF57: the claim that conformance reduces cost has no before-and-after measurement on any realigned project
- undated · CCF56: nothing shows the audit is accurate: its false findings and misses are unmeasured