What is gusset CLAUDE.md?

A development guide for Gusset, a code-analysis tool that maps how changes affect a project. It also documents the project's required testing and review workflow.

Use it to check a symbol's blast radius, run tests, verify the project on a clean machine, and understand the automated review and self-healing workflow.

It helps developers find affected code before editing and catch dependency, caching, and clean-environment problems before pushing changes.

gusset CLAUDE.md — Instructions file by latchkey-dev

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 instructions/latchkey-dev/gusset/claude-md

Clone the repo

git clone --depth 1 https://github.com/latchkey-dev/gusset

Per session 638 This file is loaded in full into every session.

When invoked 638 The same file — it is already loaded in full.

Security scan C 2 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.

Model	Per session	Once invoked
Fable 5	$0.00638	$0.00638
Opus 5	$0.00319	$0.00319
Sonnet 5	$0.00128	$0.00128
Haiku 4.5	$0.00064	$0.00064

Measured 3d ago against content hash 817d8fab2926, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade C, and why

gusset CLAUDE.md scanned grade C with 2 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 3d 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.

Downloads and executes remote codehighSupply chain

curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.

latchkey run --timeout 900 'curl -LsSf https://astral.sh/uv/install.sh | sh >/dev/null 2>&1; export PATH=$HOME/.local/bin:$PATH; uv sync -q; uv run pytest -q'

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

latchkey run --timeout 900 'curl -LsSf https://astral.sh/uv/install.sh | sh >/dev/null 2>&1; export PATH=$HOME/.local/bin:$PATH; uv sync -q; uv run pytest -q'

CLAUDE.md · 51 lines

How it starts

The opening of the file, as written. The whole thing — 51 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Gusset — development guide

Gusset develops Gusset. This repo is its own first customer: the custodian runs on our PRs (see gusset.toml), and the dev loop below is mandatory for humans and agents alike — every workflow pain point found here is a product bug worth fixing.

The dev loop

Before changing shared code, ask Gusset for the blast radius:

gusset index . --db .gusset/graph.db
gusset impact --symbol <qualname> --db .gusset/graph.db --yes
# or after editing:  gusset impact --diff HEAD --db .gusset/graph.db --yes

Use the verified impact list as the checklist of call sites to update.

Test locally: uv run pytest -q — no green, no commit.

Verify on a clean machine (catches deps/caching assumptions your working tree hides) before pushing anything substantial:

latchkey run --timeout 900 'curl -LsSf https://astral.sh/uv/install.sh | sh >/dev/null 2>&1; export PATH=$HOME/.local/bin:$PATH; uv sync -q; uv run pytest -q'

Push and let the custodian review: the Gusset Action comments the blast radius on the PR; CI runs on Latchkey runners (latchkey-small).
If CI fails and self-heal couldn't fix it: the Latchkey MCP server (setup: docs/howto/autonomous-stack.md) exposes list_failed_runs / get_failure_bundle for triage, dispatch_workflow to re-fire, and run_job for ad-hoc clean-runner commands.

House rules

The graph is the oracle: never let a workflow treat model output as truth. New workflows follow docs/howto/write-a-workflow.md and the 8-point checklist in docs/explanation/graph-engineering.md.
probe/ is the only module allowed to import pandaprobe; everything degrades to off without credentials.
No LLM code in oracle/ or graph/, ever.
evals/ results are append-only history — rerun, don't rewrite.
Every friction point in this loop goes in DOGFOOD.md — it is the product backlog the tool earns by being used, and it is public: write entries as honestly as if a stranger will read them, because one will.

Read the full file on GitHub · 51 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.

3d ago First seen · 51 lines · 638 tokens per session scan C 817d8fab2926

Subscribe to this mod's changes

gusset CLAUDE.md is an instructions file published in the GitHub repository latchkey-dev/gusset (5 stars, last pushed 9d ago), licensed MIT. It adds 638 tokens to every session, about $0.0032 per session on Opus 5. A static security scan graded it C with 2 findings (downloads and executes remote code, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Autonomous Agents Code review ai-agents code-graph developer-tools impact-analysis langgraph

laborforge-release CLAUDE.md

Claude Code instructions for EricLortie/laborforge-release, covering claude.md, first-run reality (read before setup — humans and agents), critical rules, never hardcode model names and docker-first development.

EricLortie/laborforge-release · 3,436 tokens

laborforge-release AGENTS.md

AGENTS.md instructions for EricLortie/laborforge-release, covering agents.md — laborforge, non-negotiables (you will break things otherwise), quick commands, rebuild api + workers after backend changes and no-llm smoke gate before any expensive/seed step.

EricLortie/laborforge-release · 1,100 tokens

gusset CLAUDE.md