Find code by asking what it does. A CLI for coding agents that uses Jev to discover relevant files and source context.
TypeScript ai-sdkclaude-codeclicode-searchcodexcoding-agentscontext-retrievaldeveloper-toolsjevsemantic-searchtypescriptvercel-ai-gateway
Sync skills from dzhng/skills@ad5f227
David Zhangab8b53222f9b
diff --git a/.agents/skills/audit-agents/assets/example-agents.md b/.agents/skills/audit-agents/assets/example-agents.mdindex f82bbe2..1d5e62e 100644--- a/.agents/skills/audit-agents/assets/example-agents.md+++ b/.agents/skills/audit-agents/assets/example-agents.md@@ -28,6 +28,8 @@ Write the test first. Before changing behaviour or fixing a bug, invoke <skill:A change that shouldn't alter behaviour (a refactor, a performance change) must leave the output unchanged, or be a named decision.+A failure you see is yours to fix, whether or not your change caused it. Don't hand back a codebase with failures in it; leave it better than you found it.+Never loosen a requirement to make a check pass. A narrow pass proves a narrow claim: say what you verified, what you assumed and what is unfinished.Don't wait on a long run. Start it in the background and keep working. Give it a visible sign of progress and a point where you stop, and never repeat a failure unchanged.diff --git a/.agents/skills/codex/SKILL.md b/.agents/skills/codex/SKILL.mdindex 16276db..c1a8508 100644--- a/.agents/skills/codex/SKILL.md+++ b/.agents/skills/codex/SKILL.md@@ -54,19 +54,37 @@ of eyes on — a refactor, a tricky algorithm, renderer work, a security-sensitichange — before declaring it done or committing. Skip it for trivial edits(typos, comments, doc-only).-1. Pick the diff scope: `codex review --uncommitted` for working-tree changes,- `--base <branch>` for a branch diff, `--commit <sha>` for a landed commit.- Scope flags and custom instructions are mutually exclusive (despite what- `--help` implies): `codex review "<instructions>"` reviews the default- scope with your framing, a scope flag takes no prompt. When you do write- instructions, scope the risk area — never state the answer you expect- (unprimed, same discipline as+1. Fix the diff scope before launching: `--uncommitted` for working-tree+ changes, `--base <branch>` for a branch diff, `--commit <sha>` for a landed+ commit. Resolve a landed commit to its full SHA. Use the non-interactive+ review entry point with an explicit read-only sandbox and separate progress+ and verdict files in a fresh temporary directory:++ ```sh+ review_dir=$(mktemp -d "${TMPDIR:-/tmp}/codex-review.XXXXXX")+ codex exec --sandbox read-only -C "$(git rev-parse --show-toplevel)" review \+ --commit "$(git rev-parse HEAD)" --json -o "$review_dir/verdict.md" \+ < /dev/null > "$review_dir/events.jsonl" 2> "$review_dir/stderr.log"+ ```++ Background long runs and retain their process/session handle. Scope flags+ and custom instructions are mutually exclusive: a scope flag takes no+ prompt. For custom framing, specify the exact diff to inspect in the prompt;+ don't rely on default scope or traverse all refs/checkpoint history. Never+ state the answer you expect (unprimed, same discipline as[screenshot-critique](../screenshot-critique/SKILL.md)).-2. Triage every finding: confirm it against the code before acting. Preserve+2. Monitor progress in `events.jsonl` and failures in `stderr.log`. An active+ reviewer reading relevant code is not hung merely because five minutes+ elapsed. If it stalls or drifts outside the diff and its dependencies,+ inspect the last command/error before interrupting; correct the cause or+ narrow the task before retrying. Completion requires exit status zero,+ `turn.completed`, and a nonempty final verdict. A timeout, `turn.failed`, or+ missing verdict is incomplete, never a clean review.+3. Triage every finding: confirm it against the code before acting. PreserveCodex's evidence boundaries — an inference it labelled is not a fact.Overlap with your own doubts is high-priority evidence; a finding youdismiss needs a stated reason, not silence.-3. Report the outcome to the user — what Codex flagged, what you fixed, what+4. Report the outcome to the user — what Codex flagged, what you fixed, whatyou dismissed and why. Done when every finding is either fixed orexplicitly dismissed.@@ -139,7 +157,7 @@ changes. "Process running" is NOT "working."## Rules- Don't touch the working tree while a Codex exec is running on it.-- `--sandbox read-only` (the default) for consultation and questions;+- Explicit `--sandbox read-only` for reviews, consultation and questions;`workspace-write` only for delegated implementation.- `--dangerously-bypass-approvals-and-sandbox` is reserved for tasks that mustrun browsers/servers/full suites (above) — dedicated worktree, self-authoreddiff --git a/.agents/skills/write-skills/SKILL.md b/.agents/skills/write-skills/SKILL.mdindex 48136d4..b9d4999 100644--- a/.agents/skills/write-skills/SKILL.md+++ b/.agents/skills/write-skills/SKILL.md@@ -19,9 +19,10 @@ will do.the body.2. **Spend tokens like they are scarce.** Assume the agent is already good at- general reasoning. Keep only non-obvious workflow, domain constraints,- tool choices, failure modes, and validation rules. Delete background,- motivation, and generic advice.+ general reasoning. State the direction and let smart models figure out+ implementation details. A wrapper skill can be two sentences: name the+ capability to use and the skill to follow. Add mechanics only when a+ demonstrated failure requires them.3. **Write procedures, not essays.** Prefer imperative rules, decisionpoints, and small examples. A good skill changes behavior in the next