job-apply-plugin
AGENTS.md139 starsChanged 4 months agoMITView on GitHub
Determine the repository boundary with git rev-parse --show-toplevel. Keep searches inside it; use rg and rg --files. Do not recursively search dependencies, generated output, caches, or sibling worktrees unless the task requires them. Never use find .. from a worktree. Finish the requested change, run affected checks, and fix failures caused by the change. Read documentation relevant to the changed workflow; no full repository survey is required for a small edit. Use npm run test:affected -- --base origin/staging for the…
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 32,671 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository10 other files it gives its agents.
- agent-workflowsSkill
- account-setupSkill
- answer-memorySkill
- application-setupSkill
- job-applySkill
- job-preferencesSkill
- job-searchSkill
- job-title-discoverySkill
- job-workspaceSkill
- saved-jobsSkill
The file
We split this AGENTS.md into the rules it gives an agent
4 rules. About 0 are common, 2 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 11
Use npm run test:affected -- --base origin/staging for the affected suite selection.
- Distinctive: line 24
When a local gate is red, identify whether the same failure reproduces on the clean base and report the gate as red until it passes.
- Distinctive: line 28
Keep behavioral, CLI, JSON, HTTP, privacy, and persisted-data contracts stable through the modernization.
- Very specific: line 39
Run npm run check:size for relevant changes.
≈ “Common” is an estimate: at least 1 in 5 of the 41,581 AGENTS.md 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
Repository guidance
Search hygiene
Determine the repository boundary with git rev-parse --show-toplevel. Keep searches inside it; use rg and rg --files. Do not recursively search dependencies, generated output, caches, or sibling worktrees unless the task requires them. Never use find .. from a worktree.
Task completion and verification
Finish the requested change, run affected checks, and fix failures caused by the change. Read documentation relevant to the changed workflow; no full repository survey is required for a small edit.
Use npm run test:affected -- --base origin/staging for the affected suite selection; see testing tiers for options and broader release checks. Required tiers and CI must run without CPython. Use the pinned historical checkout in testing tiers for Python comparisons. Use npm run test:migration-once for the separate historical Store migration acceptance.
Use the existing commit/push hooks rather than repeating a successful unchanged suite. Isolated fixture checks may run without repeated approval. Disruptive visible-browser/native account tests retain their explicit opt-in; never run them against the owner’s active browser merely because another local check needs them.
For skill changes, keep entry points concise and load references by task. Coordinate agent instructions with Companion status, draft preservation, and handoff behavior. Preserve manual submission, consent, managed-resume identity, and revision boundaries. Include reachable references in documentation and installed-package checks.
Buffer context is relevant only to Buffer publishing or analytics tasks; use the available buffer-workflows skill for those tasks.
Code Review Rules
- Before creating a PR, run
codex-pr-review-toolkit:review-pull-requestas a read-only branch review against the intended PR base. Resolve validated findings locally, rerun affected checks, and repeat the review if the diff changes materially; only then push/create the PR. If the toolkit is unavailable, stop and report that the pre-PR review gate was not met rather than using CI as the first review pass. - For canonical Store, profile, or Companion mutation changes, verify exact-revision conflict handling, selective writes, user provenance, and preservation of unrelated facts/preferences and browser drafts; these are contracts exercised by
npm run test:affected -- --base <intended-base>and the relevant workspace tests. - For skill, runtime, or packaging changes, verify the installed Codex/Claude artifacts and Companion handoff with
npm run test:release; source-tree tests alone do not prove the packaged workflow. - When a local gate is red, identify whether the same failure reproduces on the clean base and report the gate as red until it passes; do not present a baseline reproduction as a green result.
Codebase modernization architecture
Keep behavioral, CLI, JSON, HTTP, privacy, and persisted-data contracts stable through the modernization. Preserve human-only final submission. Python and TypeScript writers must not mutate the same live Store during migration.
Source-size policy
Scoped source, tests, test support, launchers, and checked-in runtime code must remain at or below 500 physical lines. A physical line is each newline plus a final non-empty unterminated line. Do not evade this policy through minification, statement packing, comment deletion, or generated indirection.
Run npm run check:size for relevant changes. Existing exceptions in
.source-size-baseline.json may only shrink: lower a ceiling in the same change
when its file shrinks, remove the entry once the file is at most 500 lines, and
remove obsolete entries when a file is renamed or deleted. Never add an
exception or raise a ceiling.
Worker packages
Each worker package must declare allowed_files. If it touches an oversized
file recorded in .source-size-baseline.json, its allowed_files must name at
least one extraction target for that file. Keep module dependencies directional:
new leaf modules may depend on shared primitives, but shared primitives and
facades must not depend on leaves. One owner edits the Store facade and one
owner edits the workspace bootstrap at a time.
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.

