~ / rigs / norton77930 / Provost

norton77930/Provost

Classify every request, let cheaper models do the typing, review only on risk: orchestration policy, roles and handoff runtime for Claude Code

↗ GitHub ★ 3 NOASSERTION updated 17d ago project Claude Code Claude plugin
stack: Python
share on X
ARCHETYPE
Orchestrator
A bench of specialised subagents. The main agent mostly delegates.
CONTEXT TAX · EVERY TURN
~886 tokens
Featherweight · median rig: 2.3k · breakdown
GUARDRAILS
0/5
No committed guardrails · details
Quality 0/100 Guardrails v2

Copy this rig

# review before running: this installs third-party code
$ npx degit norton77930/Provost/agents ./rig-provost/agents
$ npx degit norton77930/Provost/skills ./rig-provost/skills
$ npx degit norton77930/Provost/hooks ./rig-provost/hooks

MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.

$ npx degit norton77930/Provost/skills/handoff .claude/skills/handoff
$ npx degit norton77930/Provost/skills/verification-before-completion .claude/skills/verification-before-completion
$ curl -fsSL --create-dirs -o .claude/agents/Explore.md https://raw.githubusercontent.com/norton77930/Provost/main/agents/Explore.md
$ curl -fsSL --create-dirs -o .claude/agents/acceptance-verifier.md https://raw.githubusercontent.com/norton77930/Provost/main/agents/acceptance-verifier.md
$ curl -fsSL --create-dirs -o .claude/agents/architecture-reviewer.md https://raw.githubusercontent.<redacted>-reviewer.md
$ curl -fsSL --create-dirs -o .claude/agents/code-reviewer.md https://raw.githubusercontent.com/norton77930/Provost/main/agents/code-reviewer.md
$ curl -fsSL --create-dirs -o .claude/agents/general-purpose.md https://raw.githubusercontent.com/norton77930/Provost/main/agents/general-purpose.md
$ curl -fsSL --create-dirs -o .claude/agents/implementer-sonnet.md https://raw.githubusercontent.com/norton77930/Provost/main/agents/implementer-sonnet.md
$ curl -fsSL --create-dirs -o .claude/agents/implementer.md https://raw.githubusercontent.com/norton77930/Provost/main/agents/implementer.md
$ curl -fsSL --create-dirs -o .claude/agents/mech-executor.md https://raw.githubusercontent.com/norton77930/Provost/main/agents/mech-executor.md
$ curl -fsSL --create-dirs -o .claude/agents/test-analyst.md https://raw.githubusercontent.com/norton77930/Provost/main/agents/test-analyst.md

This rig commits no guardrails. Here is the community baseline instead — the deny/ask rules most often found across all 16.2k rigs:

{
  "permissions": {
    "deny": [
      "Read(./.env)",
      "Read(./.env.*)",
      "Read(**/.env)",
      "Read(~/.ssh/**)",
      "Bash(git push --force:*)",
      "Read(**/*.pem)",
      "Read(.env)",
      "Read(~/.aws/**)",
      "Read(**/*.key)",
      "Read(**/.env.*)",
      "Bash(git push --force*)",
      "Bash(rm -rf:*)",
      "Bash(rm -rf *)",
      "Bash(sudo *)",
      "Bash(sudo:*)",
      "Bash(rm -rf /)",
      "Bash(git push -f:*)",
      "Read(.env.*)",
      "Bash(git reset --hard*)",
      "Bash(rm -rf /*)"
    ],
    "ask": [
      "Bash(git push:*)",
      "Bash(git push *)",
      "Bash(gh pr merge *)",
      "Bash(git commit:*)",
      "Bash(rm *)",
      "Bash(gh pr merge:*)",
      "Bash(gh pr create *)",
      "Bash(git reset --hard:*)",
      "Bash(npm publish:*)",
      "Bash(git rebase *)"
    ]
  }
}

Skills (2)

Subagents (9)

Explore
model: haiku
Read-only codebase exploration and evidence gathering — locating files, tracing behavior, answering implementation questions before a plan or review. Cannot write, so it is safe to dispatch whenever r
acceptance-verifier
model: sonnet
Strict read-only acceptance check of visual, live and threshold gates — compares each gate's evidence (screenshot, command output, measurement, operation log) with its source (mockup, spec, issue) and
architecture-reviewer
model: fable
Strict read-only architecture review for a declared claim — component boundaries, interfaces, data flow, dependencies, operational risk, long-term complexity. Dispatch it when the change touches an ex
code-reviewer
model: fable
Strict read-only code review for a declared claim — correctness, regressions, security, error handling, data conversion, compatibility, missing tests. Dispatch it when an R1 or R2 risk trigger fires —
general-purpose
model: sonnet
Handles bounded multi-step tasks that do not fit a dedicated role or require tools unavailable to that role. Prefer existing specialist roles when their scope fits.
implementer-sonnet
model: sonnet
Implements one approved, bounded brief whose method is settled and whose acceptance is clear, using focused TDD on Sonnet at high effort. Not for concurrency, lifecycle/consistency, changes to how sec
implementer
model: opus
Implements one approved, bounded brief using focused TDD. Follows the main session dispatch scope and reports changed files, evidence and blockers.
mech-executor
model: haiku
Mechanical batch execution of a fully specified, repetitive change — renames, pattern application, wording or version changes, same-shaped edits across many files, docs or tests whose exact content th
test-analyst
model: haiku
Read-only test execution and regression analysis — dispatch it to run existing tests, reproduce failures, and analyze regression behavior without filling the main-session context with test output. Use

Similar rigs

copied ✓