chisel

chisel is a skill for Claude Code, Codex from IronAdamant/Chisel. It costs 77 tokens per session (927 once invoked), scanned A, original, MIT.

A test-impact analysis tool for Git projects. It maps tests to code and uses changes and history to show which tests may be affected.

In plain words
What is it for?
Use it to find tests affected by recent edits, suggest tests, identify coverage gaps and stale tests, view risk maps, and record test results. It can run with a command-line interface or connected MCP tools and stores project data in a .chisel folder.
Why use it?
It reduces guesswork about what to test after a code change and highlights risky or poorly tested areas.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/ironadamant/chisel/skills
Any agent
npx skills add IronAdamant/Chisel --skill skills
Clone the repo
git clone --depth 1 https://github.com/IronAdamant/Chisel

Made for: Claude Code, Codex.

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 chisel

README.md
[![agentmods](https://agentmods.dev/badge/skills/ironadamant/chisel/skills.svg)](https://agentmods.dev/skills/ironadamant/chisel/skills)
Your own site
<a href="https://agentmods.dev/skills/ironadamant/chisel/skills"><img src="https://agentmods.dev/badge/skills/ironadamant/chisel/skills.svg" alt="Measured on agentmods" height="20"></a>
Per session 77 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 927 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00077 $0.00927
Opus 5 $0.00039 $0.00464
Sonnet 5 $0.00015 $0.00185
Haiku 4.5 $0.00008 $0.00093

Measured 5d ago against content hash ce64a27598c2, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

chisel 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 5d 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.

skills/SKILL.md · 47 lines

How it starts

The opening of the file, as written. The whole thing — 47 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Chisel — Test Impact Analysis for Agents

Chisel (source: ~/Documents/coding_projects/Chisel, CLI: chisel, MCP server: chisel, PyPI: chisel-test-impact) maps tests to code and code to git history, answering: what to run, what's risky, who touched it. Zero dependencies, per-project .chisel/ SQLite DB.

Core loop (day-to-day)

chisel analyze .                  # first time, or after big structural changes
# ... edit code ...
chisel diff-impact                # which tests are impacted by my changes?
chisel run -- pytest tests/      # run tests AND auto-record pass/fail results
chisel triage                     # top risks + coverage gaps + stale tests in one call
chisel update                     # incremental refresh (near-instant when nothing changed)

Prefer the MCP tools when the chisel server is connected (26 tools): diff_impact, suggest_tests, triage, risk_map, test_gaps, record_result, stats; start_job/job_status/cancel_job for long analyses. Pass auto_update=true on query tools to refresh a stale DB inline. The CLI is equivalent and scripting-safe.

Key facts (v0.13–0.15+)

  • gitignore-aware: ignored trees (vendored deps, build output, fixture dirs) are never scanned or traversed; untracked-but-not-ignored files ARE visible, so working-tree analysis still works. CHISEL_INCLUDE_IGNORED=1 overrides; non-git projects are unfiltered.
  • Real CLI exit codes since v0.13: chisel analyze && pytest scripting works (status: error|git_error → 1).
  • No-op update is near-instant (edge_rebuild_skipped: true); a one-file update rebuilds in seconds, not minutes.
  • Uncommitted files: pass working_tree=true (CLI: --working-tree) to risk_map / test_gaps / suggest_tests / diff_impact / triage.
  • CLI defaults match engine for risk-map: line-weighted coverage + proximity on; --auto-update on the same tools that support MCP auto_update.
  • Monorepos: CHISEL_SHARDS=pkg1,pkg2 (or .chisel/shards.toml); shard param on analyze/update/start_job (MCP and CLI --shard). Query tools aggregate across shards; cross-shard test edges are not yet routed.
  • source trust order on suggestions: hybrid > direct > import_graph > co_change > static_require > working_tree > fallback.
  • Solo-author repos: co-change coupling is naturally sparse — rely on import_partners / import_coupling.
  • Risk scores: read _meta.uniform_components (and reweighted) before trusting the composite; exclude_new_file_boost=true for stable long-term rankings.
  • stale_tests: DB edges only (no working-tree mode).

Read the full file on GitHub · 47 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. 5d ago First seen · 47 lines · 77 tokens per session scan A ce64a27598c2

Subscribe to this mod's changes

chisel is a skill published in the GitHub repository IronAdamant/Chisel (2 stars, last pushed 1mo ago), licensed MIT. It adds 77 tokens to every session and 927 once invoked, about $0.0004 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.

Related

Other skills, from other repositories

codesynapse-cli

Use this skill to answer codebase architecture questions, trace dependencies, find callers/callees, calculate blast radius, and explore the code graph. ACTIVATE when the user asks "how does X work", "what handles Y", "where is Z defined", "explain the mechanism", "who calls", "blast radius", "shortest path", "trace…

sohilladhani/codesynapse · 182 tokens

wikipedia-glossary

Enrich an Open Mind project's glossary with standard term definitions from Wikipedia. For each already-extracted term, look up its authoritative Wikipedia definition, tidy it, and write it back to the glossary as a separate, attributed field; terms with no confident match keep their verbatim in-project definition. Use…

HelloThisWorld/open-mind · 91 tokens

fmm

MCP-first code navigation for this codebase. Use before any symbol lookup, file search, dependency trace, impact analysis, or codebase evaluation — replaces grep/glob/read with O(1) fmm tool calls. Trigger when: starting any task involving unfamiliar code, navigating code structure, finding where a symbol is defined…

srobinson/fmm · 96 tokens

glossary

Extract a project's term/acronym definitions VERBATIM from its own authoritative sources, each with file:line provenance and a grounded usage profile; resolve a term deterministically or return an honest "not found" — never generated.

HelloThisWorld/open-mind · 50 tokens

capability-router

Route a request to exactly one Open Mind capability (glossary / structure / search). A local model may decide when ready, but its choice is validated against the known capability set and falls back to a deterministic if-else floor — the model never invents or overrides a capability unchecked.

HelloThisWorld/open-mind · 62 tokens

code-graphs

Build deterministic structure / dependency / call / entry-point-flow graphs from a repository's real code, with source-traceable file nodes and definitions. No invented nodes or edges; ambiguous references are flagged, not guessed.

HelloThisWorld/open-mind · 47 tokens