agentleFS
hushh-labshushh-labs/hushh-research/.agents/skills/reviewer-app-testing/SKILL.md

reviewer-app-testing

Skill40 starsChanged 4 days agoApache-2.0View on GitHub

Use when operating reviewer-account browser rehearsals that require BYOK unlock, memory-only decrypted information, and same-session client navigation.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Avoid overlap with security-audit, pkm-upgrade-rehearsal, and frontend-cache-coherence.

    line 11
  2. Distinctive:

    Run the executable preflight before Chromium.

    line 57

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Hussh Reviewer App Testing Skill

Purpose and Trigger

  • Primary scope: reviewer-app-testing
  • Trigger on reviewer Playwright flows, BYOK unlock continuity, protected route chains, or cold-session re-unlock proof.
  • Avoid overlap with security-audit, pkm-upgrade-rehearsal, and frontend-cache-coherence.

Coverage and Ownership

  • Role: spoke
  • Owner family: security-audit

Owned repo surfaces:

  1. .agents/skills/reviewer-app-testing
  2. hushh-webapp/scripts/testing/reviewer-test-identity.mjs
  3. hushh-webapp/scripts/testing/verify-signed-in-routes.mjs

Non-owned surfaces:

  1. consent-protocol/hushh_mcp/vault
  2. hushh-webapp/lib/pkm
  3. scripts/ci/pkm-upgrade-gate.sh

Do Use

  1. Canonical reviewer identity, authentication, passphrase-wrapper unlock, and vault-key integrity proof.
  2. Same-browser-context protected flows where decrypted information and keys must remain memory-only.
  3. Separate cold-context reauthentication and re-unlock proof after the same-session route chain.
  4. Shared Playwright helpers used by domain-specific reviewer rehearsals.

Do Not Use

  1. Vault cryptography or PKM storage implementation changes; use security-audit.
  2. PKM restructuring, preservation, scope, rollback, or payload acceptance; use pkm-upgrade-rehearsal.
  3. Generic public-route browser checks that do not depend on an unlocked vault.
  4. Shared reviewer mutation without explicit operator authority for the current task.

Read First

  1. .agents/skills/reviewer-app-testing/references/byok-reviewer-browser-contract.md
  2. .agents/skills/reviewer-app-testing/references/reviewer-rehearsal-preflight.md
  3. hushh-webapp/scripts/testing/reviewer-test-identity.mjs
  4. hushh-webapp/lib/testing/native-test.ts
  5. hushh-webapp/components/app-ui/native-test-bootstrap.tsx
  6. hushh-webapp/components/app-ui/native-test-router.tsx
  7. hushh-webapp/lib/utils/browser-navigation.ts

Workflow

  1. Classify the run as read-only or explicitly mutation-authorized; ordinary review is read-only.
  2. Run the executable preflight before Chromium; it resolves the canonical reviewer without printing secrets and proves review mode is enabled.
  3. For local runs, use scripts/env/reviewer_mode.sh; restart the backend after enabling and disabling it.
  4. Let the default network guard block unapproved state-changing requests. A block is a failure, never a reason to weaken the guard.
  5. Prove a cold authenticated entry visibly hard-gates on the vault before supplying any passphrase.
  6. Unlock, verify vaultKeyHash, then use Next client navigation and assert vault_unlocked after every protected transition.
  7. Fail on critical API 5xx, a blocked mutation, identity mismatch, or any loss of vault continuity.
  8. Close the context, then prove cold-session authentication and re-unlock separately.
  9. For Agent Chat changes, run verify-reviewer-agent-chat.mjs with explicit REVIEWER_ALLOW_SHARED_MUTATIONS=true; it must complete a real prompt turn and prove consumer-safe errors, profile-only avatars, idle-status removal, and overflow safety.
  10. Trusted Devices: verify-reviewer-trusted-devices.mjs (read-only). Always compose createReviewerSessionHarness; never hand-roll one (see the preflight reference).
  11. Wallet, consent lifecycle from chat, large Memory imports: verify-reviewer-wallet.mjs, verify-reviewer-consent-chat.mjs, verify-reviewer-memory-import.mjs (mutation-authorized; the preflight reference states what each proves).
  12. Physical native reviewers hand off to mobile-parity-audit and its references/physical-reviewer-preflight.md; browser authorization does not enable XCTest or unlock a device. Local enablement and identity reconciliation follow .agents/skills/reviewer-app-testing/references/reviewer-rehearsal-preflight.md.

Handoff Rules

  1. Vault and encrypted-storage implementation work routes to security-audit.
  2. PKM upgrade and exact payload rehearsal routes to pkm-upgrade-rehearsal.
  3. Cache behavior proof routes to frontend-cache-coherence; broad security work returns to security-audit.

Required Checks

./scripts/ci/reviewer-app-testing-check.sh
./bin/hushh agents route-task reviewer-app-rehearsal --text
node .agents/skills/reviewer-app-testing/scripts/reviewer-rehearsal-preflight.mjs --help
consent-protocol/.venv/bin/python .agents/skills/reviewer-app-testing/scripts/reconcile-reviewer-identity.py --help

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.