szl-figure-data-contract
Renders and replays a bounded local CSV-to-SVG scatter or line figure with row IDs, units, numeric caption assertions and byte receipts.
Skill by szl-holdings
Chalk Labs
Need help deploying szl-figure-data-contract?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/szl-holdings-szl-figure-data-contract
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: documents, spreadsheets, design, data
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md and code. Checked Nov 1, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
- Runs as
- Python package
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/szl-figure-data-contract
What's inside
- Skill name
szl-figure-data-contract- Description
- Renders and replays a bounded local CSV-to-SVG scatter or line figure with row IDs, units, numeric caption assertions and byte receipts. Use to inspect a declared figure outside its authoring application and detect dropped, reordered or altered points.
- License
- Apache-2.0
- Instructions
- 52 lines in SKILL.md · view on GitHub
Before installing
- Ships executable scripts
Files (4)
- SKILL.md
- data.csv · assets/
- spec.json · assets/
- run.py · scripts/
SKILL.md (first lines)
# Figure data contract Use Python 3.10+ and the standard library. From this skill directory: ```bash python -I -B scripts/run.py render assets/spec.json --root . --output figure-bundle python -I -B scripts/run.py verify figure-bundle --root . ``` The specification in `assets/spec.json` names a local UTF-8 CSV, distinct ID/x/y columns, axis units, scatter or line geometry, and count/minimum/maximum caption assertions. Decimal observations and extrema are strings, shown in fixed notation (10000, not 1E+4). An axis may name `relation_column` and `unit_column`; each row must then hold `=` and the axis unit. Censored or other-unit rows, and undeclared relation, qualifier or unit(s) columns, are `UNKNOWN`; units are never converted. Line points retain CSV row order; no sorting, aggregation, interpolation or hidden filtering occurs. The synthetic example uses `assets/data.csv`. Rendering refuses incorrect caption assertions before writing a bundle. Output must be a new relative directory with an existing parent under the selected root. It contains an SVG figure, plotted-data JSON and a receipt with hashes of input, specification, outputs and this helper. Verification recomputes the declared data, parses SVG point geometry and visible units/caption, and checks the byte bindings. Only canonical generated SVG bytes pass; declarations, external stylesheet instructions, added elements and otherwise equivalent reformattings are refused. It does not execute a supplied notebook, script or arbitrary transform. `MATCH` (exit 0) means this declared mapping and numeric caption agree with the retained bytes; `MISMATCH` (exit 1) means a binding or checked quantity differs; `UNKNOWN` (exit 2) means inputs are invalid, unsupported, missing or unavailable. Readable retained inputs with changed byte hashes are `MISMATCH` even if the changed contents would no longer parse; missing or unreadable files stay `UNKNOWN`. The verifier shares decimal parsing and coordinate arithmetic with the renderer; its separate SVG inspection is not an independently authored numerical method. Keep original source bytes with the bundle so another reviewer can rerun it. Hashes are unsigned: a person can alter all inputs and create a new internally consistent receipt. The skill does not establish measurement authenticity, scientific correctness, independent replication or arbitrary existing figure pixels. Raw CSV is not copied into the bundle, but plotted coordinates, row IDs and summary values can disclose study data; review them before sharing. Bounds: CSV 256 KiB, 10,000 rows, 64 columns; JSON spec 16 KiB; unique safe row IDs; finite decimals with at most 32 digits within 1e-100 to 1e100; SVG/sidecar reads 2 MiB each. Paths must be relative POSIX paths. Symlinks, junctions, reparse points, traversal, special files and credential-like paths are refused. This assumes a cooperative local filesystem; it is not a race-safe sandbox. No network, credentials or packages are needed. No RO-Crate conformance, Claude Science registration or measured agent effectiveness is claimed. For broader source/environment retention use szl-reproducibility-capsule. For a figure caption's location in a PDF use szl-paper-evidence-audit. This skill checks the declared CSV-to-figure mapping, which those checks do not replay.
From szl-holdings/szl-skills, Apache-2.0.
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 1
- Last commit (this folder)
- Oct 3, 2026 (1 month ago)
- Last release
- Oct 2, 2026 (1 month ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- Apache-2.0
- Repository
- github.com/szl-holdings/szl-skills
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get szl-figure-data-contract running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
szl-figure-data-contract skill. Agent Mart. https://agentmart.dev/skills/szl-holdings-szl-figure-data-contract (last updated 2026-10-24)