houserules-step
Skill2 starsChanged 21 days agoMITView on GitHub
Use around every change to this repository, in any workflow: read the rules for the files before you change them, then audit the change and validate its report.
In this repository12 other files it gives its agents.
- AGENTS.mdAGENTS.md
- template/AGENTS.mdAGENTS.md
- tests/AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- template/CLAUDE.mdCLAUDE.md
- designingSkill
- finishing-a-featureSkill
- migrating-knowledgeSkill
- orchestratingSkill
- planningSkill
- project-knowledgeSkill
- readme-reworkSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 10
Read the rules: Run tools/houserules-step.sh before <path>...
- Distinctive: line 11
Implement the change: Follow the rules you read.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
The houserules step
A change takes four steps. Two of them call tools/houserules-step.sh. The script writes its output to files below .houserules/step/, prints the path of each file, and prints each exit status, so you read the output with your file-read tool. Paths and refs you pass to the script are relative to the install root.
- Read the rules. Run
tools/houserules-step.sh before <path>...with every file you plan to change. The script runshouserules forand writes the rule package to.houserules/step/rules.json. Read the file. Runhouserules get <id>for a rule you are unsure about. - Implement the change. Follow the rules you read.
- Audit the change. Commit the change first:
auditjudges the commits from<base>toHEAD, and a runner check runs a project tool on the working tree. When a file that the tool reads differs fromHEAD,auditexits 2 with no rows and a line that names the file. Runtools/houserules-step.sh after <base>with the ref you started from. The script runshouserules audit --base <base>and writes the result to.houserules/step/audit.json. Read the file. Fix everyfailrow and every line inreferences.failures, then run the step again. A line inreferences.failuresis a work-item reference that fails, the line that asks for a tracker, or the line of a tracker that cannot be reached.summary.reference_failcounts them. - Write and validate the report. Write the report in the report format of
.claude/schemas/deliverables.json. Run each test and live run throughhouserules capture --out <file> -- <command>, with<file>in the workspace (default.houserules/work/), never in.houserules/step/, which the script overwrites; write each run as{ "capture": "<path>" }. Runtools/houserules-step.sh after <base> <report>. The script runs the audit with--report, thenhouserules report-facts, which fills the report's facts, thenhouserules validate, and writes their output to.houserules/step/. Fix every error.
init adds .houserules/ to .gitignore: this script writes its scratch output to .houserules/step/, and .houserules/work/, the default workspaces location, holds the workspaces of the project, which stay out of git.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

