By hrmtz
Defense-in-depth hooks for Claude Code: credential leak prevention, dangerous bash command guard, admission-keyword workflow reminders.
Operational harness for Claude Code, distilled from 6+ months of self-hosted production use.
Why a "harness"? LLM agents fail in patterned ways — credential leaks, recovery loops, premature script generation. Behavioral rules in
CLAUDE.mdcatch some of these, but the same incidents repeat. This marketplace ships structural fixes (hooks, guards, reflexive context injection) that don't rely on the agent remembering.
| Plugin | What it does | Trigger |
|---|---|---|
| harness-core | 3 hooks: credential value scrub (PostToolUse) + dangerous bash guard (PreToolUse) + admission-keyword workflow reminder (UserPromptSubmit) | Every Bash call + every user prompt |
More plugins are planned (formation skill for long-running tmux pane workers, CLAUDE.md persona templates) — see docs/ROADMAP.md once published.
# in Claude Code
/plugin marketplace add github:hrmtz/claude-harness
/plugin install harness-core@claude-harness
After install, Claude Code auto-wires the hooks via ${CLAUDE_PLUGIN_ROOT}/hooks/hooks.json. No manual ~/.claude/settings.json edit needed.
# trigger credential scrub: paste a fake key into a Bash command output
echo 'sk-ant-api03-FAKE_KEY_FOR_TEST_xxxxxxxxxxxxxxxxxxxx'
# expected: hook detects pattern, sanitizes active session jsonl, emits warning
# trigger bash guard: try a forbidden pattern
sops -d secrets.enc.yaml
# expected: PreToolUse blocks with explanation
If you'd rather copy the files directly:
git clone https://github.com/hrmtz/claude-harness
cp -r claude-harness/plugins/harness-core/hooks/* ~/.claude/hooks/
# then add to ~/.claude/settings.json — see plugins/harness-core/README.md
This marketplace is one half of a larger system. The other half is the philosophy + memory + persona doc:
docs/CLAUDE_HARNESS_DISTILLED.md — full design rationale (3-tier memory, 真田/松岡/仗助 persona stack, SOPS 2-command rule, 8 incident timeline → structural fix)Read that first if you want to understand why these hooks exist before installing them.
harness-core (this commit) — production-tested locallyharness-formation — pending public release of njslyr7 (tmux pane peer-worker daemon)harness-claude-md-template — paste-able CLAUDE.md skeletonMIT — see LICENSE.
Executes bash commands
Hook triggers when Bash tool is used
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
npx claudepluginhub hrmtz/claude-harness --plugin harness-coreCore skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Superpowers Plus core skills library for Claude Code: planning, execution routing, TDD, debugging, and collaboration workflows
Claude harness - A harness for solo developers (Vibecoders) to handle full-cycle contract development.
Intelligent prompt optimization: injects the right context at the right moment so Claude lands a better first output. Clarifies vague prompts with research-based questions, plus targeted nudges for approach selection, plan readability, workflow routing, background execution, subagent routing, output readability, user-decision questions, and plan-mode assessment