neohaskell/docs/archive/2026-07-ai-artifacts/claude-skills/integration-pipeline-preview/05-performance-design/01-hot-path-analysis

ARCHIVAL KNOWLEDGE — DO NOT USE OR REFERENCE. Superseded by codemap/ + root AGENTS.md. Scheduled for deletion once the pipeline is verified (Phase 6). Manifest: docs/archive/2026-07-ai-artifacts/MANIFEST.md

name: 01-hot-path-analysis description: Analyses the design draft against the performance methodology to surface raw hot-path findings. kind: leaf executor: opus model: claude-opus-4-7

Hot-Path Analysis

Reads the design draft and asks every section of the performance methodology against it, emitting raw findings without filtering.

Review posture

Assume the artefact under review was produced by a language model (ChatGPT-class output). Treat plausible-looking claims as unverified, expect hallucinated APIs and missed constraints, and refuse to pass anything not directly traceable to the methodology in ../../references/performance-methodology.md. Strict review is the default — benefit of the doubt goes to the rubric and to the source, never to the producer.

Inputs

• .integration-pipeline/integration-design.md — the design draft produced in phase 3.
• ../../references/performance-methodology.md — methodology digest.
• ../../references/nhcore-context.md — nhcore performance context.

Plan

1. Read the design draft → verify: file exists and is non-empty.
2. Read the performance methodology and nhcore context → verify: both load.
3. For each numbered section of the methodology, ask its question against the design draft → verify: every section produces at least an explicit "not applicable" finding or a real finding.
4. Emit a JSON array of raw findings on stdout → verify: each entry has severity, rule, location, recommendation.

Assumptions:

• No filtering happens here — that is the grounding step's job.
• The 50k req/s target frames every severity decision (Blocking / Advisory / Informational).

If any assumption fails, refuse — do not guess.

Steps

1. Load the design draft from .integration-pipeline/integration-design.md.
2. Load the performance methodology and nhcore context references.
3. For each methodology section (1-9), reason about the design draft section that maps to it; produce one or more findings.
4. For each finding, populate:
   • severity: Blocking / Advisory / Informational.
   • rule: short slug, e.g. inline-pragmas, tojson-without-toencoding, tvar-map-contention.
   • location: design draft section name or heading.
   • recommendation: a one-line fix.
5. Emit the JSON array on stdout.

Output

JSON array of raw performance findings on stdout.

Refusals

• .integration-pipeline/integration-design.md missing → refuse: "no design draft; run phase 03 first".
• Reference files missing → refuse: "performance methodology or nhcore context missing".