frontier-readiness: gap #5 audit — AGENTS.md classified (both-coupled)#180
frontier-readiness: gap #5 audit — AGENTS.md classified (both-coupled)#180
Conversation
Second file audited in the factory-vs-Zeta separation audit (gap #5). AGENTS.md is the universal onboarding handbook. Overall classification: both (coupled) — mostly factory- generic in shape, with Zeta-library-specific content in 6 areas: 1. Three load-bearing values #2 (Z-set / operator laws) 2. "What we borrow" section (DBSP / Arrow / etc. list) 3. Build and test gate (dotnet commands) 4. Code style (F# / .NET specifics) 5. Inline ZSet / algebra examples (audit on-touch) 6. Pre-v1 Status declaration (project-specific shape) Estimated refactor effort: M (more surgical edits than CLAUDE.md's S, but each isolated). Post-split location: Frontier (authoritative onboarding template); adopters fork + customise example sections. Zeta-specific content extracted to Zeta repo's own CONTRIBUTING.md (or equivalent). Section-by-section breakdown for 15 sections documented. Attribution: Otto (loop-agent PM hat). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Continues gap #5 cadence (1-2 files per dedicated tick). PR #179 (seed) merged; audit substrate now on main. AGENTS.md classified both (coupled): 15 sections broken down, 6 surgical refactor edits flagged, effort M. Post-split location: Frontier as authoritative onboarding template. PR #180 armed for auto-merge. Hypothesis from first two audits: onboarding substrate is uniformly "both (coupled)" — mostly factory-shape with embedded Zeta-examples. If holds for next several files, refactor scope becomes mechanical. Attribution: Otto (loop-agent PM hat). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 90aa169d1d
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| | "Agent operational practices" | factory-generic | Discipline around skills / agents / MCP. Generic. | | ||
| | "Build and test gate" | zeta-library-specific | `dotnet build -c Release` / `dotnet test Zeta.sln` / `TreatWarningsAsErrors` / `Directory.Build.props`. Entirely Zeta-library-specific. In Frontier, this section removes the specifics and declares "your project has a build-test gate that you run every change; fill in specifics per your language/runtime." | | ||
| | "Code style and conventions (short form)" | **both (coupled)** | F#-first / C#-wrapper / struct-tuples / Span<T> / etc. are Zeta-specific. The generic principles (clear naming / small functions / consistent style) transfer. Frontier version keeps generic principles; drops F#/.NET specifics. | | ||
| | "PR / commit discipline" | factory-generic | Small PRs / clear commit messages / Co-Authored-By. Generic. | |
There was a problem hiding this comment.
These rows describe directives that are not present in AGENTS.md (e.g., "MCP", "struct-tuples", Span<T>, and "Co-Authored-By"); the referenced sections actually contain different guidance. Because this document is used to drive the Frontier/Zeta split plan, invented inputs make the classification and refactor checklist less reliable and can trigger unnecessary or mis-scoped edits. Please align the section notes to the actual text in AGENTS.md before marking this file audited.
Useful? React with 👍 / 👎.
There was a problem hiding this comment.
Pull request overview
Extends the Frontier-vs-Zeta separation audit (gap #5) by classifying AGENTS.md and recording refactor notes needed before the multi-repo split.
Changes:
- Moves AGENTS.md from “to audit” into “audited” and adds a full audit section.
- Adds section-by-section classification, refactor notes, and rationale for AGENTS.md.
| the authoritative template; adopters fork + customise the | ||
| example sections. | ||
|
|
||
| **Length:** 359 lines. |
There was a problem hiding this comment.
The recorded length for AGENTS.md is incorrect. The repo currently has AGENTS.md at 360 lines, so this should be updated to match (or rephrased to avoid hardcoding a line count that will drift).
| **Length:** 359 lines. | |
| **Length:** substantial; treat the exact line count as | |
| volatile as `AGENTS.md` evolves. |
| ## Audit — AGENTS.md | ||
|
|
||
| **Overall classification:** **both (coupled)** — the file is | ||
| explicitly designed as a *universal onboarding handbook* | ||
| ("works with any AI harness ... as well as for human | ||
| contributors") which is factory-generic in shape, but | ||
| contains several Zeta-library-specific sections that need | ||
| surgical edits before the split. | ||
|
|
There was a problem hiding this comment.
The audit schema earlier in this doc includes an "Audit date + auditor" field for each audited file, but this AGENTS.md audit block doesn't record either. Please add an explicit audit date (UTC) and role-based auditor identifier (no personal names) for traceability, or adjust the schema if this information is intentionally omitted.
| ### Files audited | ||
|
|
||
| - **CLAUDE.md** (below, this seed fire) | ||
| - **AGENTS.md** (below, Otto-9 fire) |
There was a problem hiding this comment.
Repo convention forbids contributor name attribution in docs; this entry includes a direct name ("Otto-9") instead of a role-ref. Please replace with a role-based reference (e.g., "loop-agent PM" / "auditing agent") or a neutral fire-id, and avoid embedding personal names in this doc (see docs/AGENT-BEST-PRACTICES.md operational rule).
Summary
Section-by-section classification
15 sections classified:
dotnetcommands)Refactor notes
Before split, 6 surgical edits:
Effort: M (more edits than CLAUDE.md's S; each isolated).
Post-split location
Frontier hosts AGENTS.md as the authoritative onboarding template; adopters fork and fill in their substrate-specific sections. Zeta retains extracted Zeta-specific content in its own CONTRIBUTING.md.
Test plan
Attribution
Otto (loop-agent PM hat); continuing gap #5 cadence (1-2 files per dedicated tick).
🤖 Generated with Claude Code