01 · Roasts
Commit firehose, heatmap blackout
4,190 yearly commits arrived mostly in a late sprint; the first 43 heatmap weeks are blank enough to make a calendar nervous.
Quality outran applause
release-policy has named consumers and au-fpa-pack has a 90% coverage gate, yet the whole account has 1 star.
Python accounting multiverse
Seven accounting engines, FP&A, tax corpus, review controls, and planning cubes: apparently spreadsheets were too easy.
CI has trust issues
Pinned actions, audits, schema contracts, Gitleaks, checksums, and browser tests suggest every comma is considered an insider threat.
Built using
Zoral
Shadows one worker for a week, then takes over their job with zero extra setup. Behaves exactly like the original.
zoral.ai
02 · Category breakdown
- Impact25% weight68C
- Consistency20% weight85A
- Quality20% weight77B
- Depth15% weight65C
- Breadth10% weight55D
- Community10% weight30F
03 · Stats
365-day commit heatmap
54 active days
Language distribution
- Python81%
- HTML7%
- JavaScript6%
- Shell3%
- PowerShell2%
- Power Query1%
04 · Numbers
Owned repos
non-fork
15
Commits
last 12 months
4,190
Followers
9
Joined GitHub
Dec 2023
05 · Top repos
ryanduguid /
au-fpa-pack
A substantial Australian FP&A Python package with Xero AU parsing, GST/BAS timing, payroll, cash forecasting, Excel export, CLI workspace tooling, synthetic examples, and unusually rigorous CI/test gates despite zero stars.
ryanduguid /
australian-accounting
A documented, locally runnable Australian accounting toolkit spanning seven Python engines and an MCP server, with strong safety boundaries, schema-focused tests, CI, and published-package usage examples despite limited visible adoption.
ryanduguid /
ryanduguid.github.io
A substantial, documented Australian accounting static site with nine calculators, 100 question guides, provenance-controlled rate data, and unusually rigorous automated, browser, SEO, accessibility, and CI checks; adoption remains unproven at 0 stars.
ryanduguid /
llm-tax-guardrails
A carefully engineered Australian tax-LLM guardrail guide with deterministic generation, integrity checks, an inert evaluation harness, 45 behaviour cases, and strong CI, but no demonstrated adoption beyond the repository.
ryanduguid /
release-policy
A documented, tested policy engine with reusable GitHub release workflows, deterministic archives, provenance-oriented manifests, canary auditing, and fail-closed egress checks; adoption is evidenced by named consumer repositories but public traction is still zero-star.
ryanduguid /
accounting-review-pipeline
A documented, tested, CI-backed Python accounting pipeline with seven independently versioned review components, deterministic Decimal controls, reconciliation, readiness gates, and offline conformance artifacts; adoption is not yet evidenced by its 0 stars/forks.
ryanduguid /
github-agent-skills
A focused, documented Python repository shipping five portable agent skills, with strict synchronization/public-safety validation, comprehensive unit tests, recorded scenario runs, and multi-platform CI.
ryanduguid /
au-tax-legislation-corpus
A substantial Python corpus builder and review-queue package with provenance-focused capture, strict contract validation, extensive tests, and multi-platform CI, but currently showing no stars, forks, or explicit external adoption.
ryanduguid /
planning-analytics-model
A polished, published PaciolisCube Python package that models IBM Planning Analytics offline, with a real rule parser/evaluator, five-cube example model, strong tests, CI, and documented accounting assumptions, but no demonstrated adoption beyond 0 stars.
ryanduguid /
australian-accounting-skills
A well-documented, safety-focused Australian accounting workflow pack with 60 distributable skills, extensive validation, and CI; adoption remains minimal at 1 star and 1 fork.
ryanduguid /
Metanoia
A carefully documented, CI-backed Python maintenance repository for profile links, identity records and portfolio provenance, with strong policy tests but limited standalone product impact and no typed application surface.
ryanduguid /
.github
A focused account-level GitHub community-health repository with detailed privacy-aware templates, documented contribution and security policies, and schema-validation CI, but no tests or evidence of external adoption.
06 · Timeline
- Dec 2, 2023Joined GitHub
- Aug 1, 2026Created australian-accounting-skills — Preparation and review workflows for Australian public practice and contracting: BAS, month-end close, payroll and workpaper tie-outs. No lodgement or advice.
- Aug 1, 2026Created Metanoia — Portfolio documentation, maintenance notes and profile tooling for the ryanduguid repositories.
- Aug 4, 2026Created au-tax-legislation-corpus — Code for building a corpus of in-force Commonwealth tax legislation, with source and provenance records.
- Aug 8, 2026Created accounting-review-pipeline — Runs deterministic month-end close checks on Xero trial-balance exports and produces exception packs for human sign-off.
- Aug 16, 2026Created llm-tax-guardrails — APES 110 and TPB Code controls, DrDebits refusal patterns and evaluation fixtures for firms using LLMs in tax work.
- Aug 17, 2026Created release-policy — Reusable workflows for releases, provenance attestations and Python package checks.
- Aug 19, 2026Created au-fpa-pack — Australian FP&A tools for openfpa (Jeff Brines / Guiderail), with 30 June years, Xero AU and GST/BAS.
- Aug 21, 2026Created australian-accounting — Local engines for ATO small-business benchmarks, Payday Super 2026 and Division 7A loan review, available through the aus-accounting-mcp server. Synthetic fixtures only.
- Aug 22, 2026Created ryanduguid.github.io — Open-source engines, MCP servers and AI agent workflows for Australian tax and accounting.
- Aug 23, 2026Created planning-analytics-model — An IBM Planning Analytics budgeting model published as source, with an offline engine that computes and tests it.
- Aug 31, 2026Created github-agent-skills — GitHub workflow skills for Codex and Claude Code that require inspected evidence.
- Aug 31, 2026Created .github — Default community health files for ryanduguid repositories
- Sep 24, 2026Most recent push to australian-accounting
07 · Compare
08 · Rubric
How this score was produced
Overall = Σ (category × weight) + gentle top-end curve
Tier thresholds
▸ How the pipeline works
- 01Scrape.Pull every non-fork repo pushed in the last 90 days, plus your contribution calendar, followers, and language byte counts — straight from GitHub's REST & GraphQL APIs.
- 02Triage.A small model reads every repo's file tree + README and picks the 20 files per repo that actually reveal how you code.
- 03Grade each repo. All repos run in parallel through a fast scoring model that reads the picked files and rates each one independently on Impact, Quality, and Depth — with evidence citations.
- 04Aggregate. A larger reasoning model combines the per-repo scores with server-computed stats (heatmap, commit cadence, language entropy, follower count) to produce the 6-dimension profile score + roasts.
- 05Correct.Deterministic server-side checks enforce anchor-scale floors (e.g. a profile with 2,000+ public commits can't score 30 Consistency) and recompute the final verdict.
~90 seconds per profile, ~$0.25 in compute. Total of ~240 files read across your top-12 repos. One rating per GitHub account per day.
▸ Data sources & caveats
- Heatmap & commit totals: GitHub GraphQL
contributionsCollection— covers the last 365 days, includes private repos when the user has opted in (default). - Language %: byte totals across the top 30 owned non-fork repos.
- Curve: a small upward nudge centered on raw score ≈ 70, capping at 100. Prevents specialists from being unfairly penalised for narrow breadth.
- Anchor corrections: when server-measured signals (e.g. privateWorkLikely, multiRepoVolume, follower count) mandate a minimum category score, the aggregation step enforces it. These are signal-conditional, not identity-based floors.