Seshat-BI: Skill for Claude Code

.claude/skills/grain-confidence-reviewer/SKILL.md

grain-confidence-reviewer is a skill for Claude Code from Kemetra/Seshat-BI. It costs 175 tokens per session (2,572 once invoked), scanned A, original, Apache-2.0.

A reviewer aid for checking whether a table's declared grain—the level of detail represented by each row—is actually unique, and for comparing two source-map.yaml versions.

In plain words
What is it for?
Use it before reviewing or approving a table's mapping, when someone asks whether its grain is confident or what changed between mapping versions.
Why use it?
It puts the measured uniqueness evidence and mapping changes in front of a human reviewer without inventing a confidence score or making the approval decision.

Skill for Claude Code

Written for Claude Code: installed under .claude/. Also seen: reads .claude/ paths.

This is Kemetra/Seshat-BI's own configuration. It tells Claude Code how to work on Seshat-BI itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything Seshat-BI configures →

Reuse

Borrowing it

Nothing to install: this file belongs to Kemetra/Seshat-BI. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/Kemetra/Seshat-BI/main/.claude/skills/grain-confidence-reviewer/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/Kemetra/Seshat-BI

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for grain-confidence-reviewer

README.md
[![agentmods](https://agentmods.dev/badge/skills/kemetra/seshat-bi/grain-confidence-reviewer/github.svg)](https://agentmods.dev/skills/kemetra/seshat-bi/grain-confidence-reviewer)
Your own site
<a href="https://agentmods.dev/skills/kemetra/seshat-bi/grain-confidence-reviewer"><img src="https://agentmods.dev/badge/skills/kemetra/seshat-bi/grain-confidence-reviewer/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for grain-confidence-reviewer

Your own site · 80×15
<a href="https://agentmods.dev/skills/kemetra/seshat-bi/grain-confidence-reviewer"><img src="https://agentmods.dev/badge/skills/kemetra/seshat-bi/grain-confidence-reviewer.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 175 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,572 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00175 $0.02572
Opus 5 $0.00088 $0.01286
Sonnet 5 $0.00035 $0.00514
Haiku 4.5 $0.00017 $0.00257

Measured 12d ago against content hash 78982a3aaa0c, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-12, from the pricing page.

Security

Grade A, and why

grain-confidence-reviewer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/skills/grain-confidence-reviewer/SKILL.md · 179 lines

How it starts

The opening of the file, as written. The whole thing — 179 lines — stays where its author put it; the contents beside it link to each section on GitHub.

grain-confidence-reviewer

The Mapping Ready gate turns on one load-bearing fact: is the declared grain actually unique on the data? That fact is already measured (profile.py's PkProof) and the diff inputs are two committed source-map.yaml versions. This skill SURFACES both for a reviewer -- a grain-confidence card and a semantic mapping diff -- and STOPS at the human seam. It deepens the shipped mapping gate (roadmap Layer 3); it adds no new gate, no new Python, no numeric score.

Scope boundary (read first)

  • Surfaces and STOPS. It RENDERS evidence (a card, a diff) for a human reviewer; it does NOT write an approval, does NOT write Gate status: CLEARED, does NOT edit source-map.yaml to clear a finding, and does NOT pick a new candidate PK or widen a grain on the human's behalf.
  • "Confidence" is evidence + status + blockers, NEVER a number. No 0.87, no auto high/medium/low label. Grain confidence is reported as exactly ONE of the four readiness statuses (not_started | blocked | warning | pass) with the measured signal cited as evidence and any blockers attached (roadmap rule #9; "No fake confidence", readiness-model.md).
  • Reuses the measured signal; re-implements nothing. It reads PkProof (total, distinct_pk, null_pk, is_unique); it never writes its own uniqueness query.
  • Judgment calls are the human's (Principle V): grain ambiguity, PII publish-safety, business rollups -- surfaced, never auto-resolved.
  • Cite Principles IV (mapping gate), V (human seam), VII (generic), VIII (live deferred).
  • ASCII only, UTF-8 no BOM.

US1 -- Grain confidence card

Read the measured signal

Read the Candidate-grain/PK numbers from mappings/<table>/source-profile.md; OR, at the deferred live boundary, re-run src/seshat/profile.py over a READ-ONLY connection (resolve_dsn + make_psycopg2_runner, the db extra). Reuse PkProof (total, distinct_pk, null_pk, is_unique) -- do NOT re-implement the uniqueness query.

Read the full file on GitHub · 179 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 12d ago First seen · 179 lines · 175 tokens per session scan A 78982a3aaa0c

Subscribe to this mod's changes

grain-confidence-reviewer is a skill published in the GitHub repository Kemetra/Seshat-BI (2 stars, last pushed 3d ago), licensed Apache-2.0. It adds 175 tokens to every session and 2,572 once invoked, about $0.0009 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.