agentmods.dev

Best Claude Code hooks

1,114 hooks in the catalogue install into Claude Code. These 4 are the ones that pass every gate: original work, not a copy · security scan A or B · a licence that permits reading the source · an actively maintained repository people actually use. Order is the composite score — reputation, freshness, safety, originality, content and traction — not anyone's opinion.

Read the table the way you would pick one: Tokens is what a hook adds to every session before you use it — forty of them at 500 each is 20,000 tokens of context you did not ask for; Scan is what 26 rules found in the file, A meaning nothing; Stars belong to the repository, so a ten-star file from a well-known vendor can sit above a copy inside a 200k-star monorepo. Every row links to a page with the install command for Claude Code, the scan reasons and what changed since last week.

← Best hooks, every agent
#ModTypeCategoryStarsScanTokensRating
1 UserPromptSubmitEverMind-AI/SkillCorpusRuns when you submit a prompt, before the agent sees it, executing hook.mjs via node. From EverMind-AI/SkillCorpus. Hook Other 608 A — —
2 PostToolUsealeclindz/seo-skill-benchRuns after a tool call finishes for Write and Edit tool calls, running an inline shell check. From aleclindz/seo-skill-bench. Hook Claude Code 51 A — —
3 StopProgrammerAnthony/Expert-Coding-HarnessRuns when the agent finishes a response, executing check-secrets.js, check-console-log.js, session-end.js and cost-tracker.js via node (4 commands). From ProgrammerAnthony/Expert-Coding-Harness. Hook Other 236 A — —
4 PreToolUsejohnkf5-ops/the-dev-squadRuns before the agent uses a tool, executing approval-gate.sh. From johnkf5-ops/the-dev-squad. Hook Security 238 A — —

Want everything, not just what passes the gates? Browse all hooks for Claude Code. This shelf recomputes nightly from the measurements.