Scope and intake
Turn a broad request into a bounded review plan with recorded ownership, constraints, and evidence goals.
22 review guides. Each one covers what to look for, what to produce and the common mistakes. Read them here or keep them as Markdown. Download all as JSON
Seven guides, in order, from scope intake through a cited handoff.
22 skills
Turn a broad request into a bounded review plan with recorded ownership, constraints, and evidence goals.
Review whether every protected operation binds the caller, tenant, resource, and action together at the decision point.
Assess session issuance, recovery, revocation, and privilege changes as one continuous account-authority lifecycle.
Review delegated identity and token flows for strict issuer, client, redirect, audience, and grant binding.
Treat each GraphQL operation, resolver, and returned field as a separate server-side authorization and disclosure decision.
Review browser code as an untrusted-data consumer: preserve origin, context, serialization, and server-authorization boundaries.
Review cache keys, response eligibility, and invalidation so a response remains correct for the requester and representation that produced it.
Keep retrieved content, model output, tool authority, and user consent separate throughout an AI-assisted workflow.
Review build and dependency trust from source selection through a verifiable artifact and its deployment authority.
Review mobile link association, client configuration, and backend APIs as complementary controls with server-owned authorization.
Review file and document processing as a chain of bounded parsing, typed validation, storage, and authorized action.
Produce a concise security record that separates observed evidence, bounded impact, uncertainty, and remediation.
Review cloud storage and service access through attributable ownership, least privilege, destination policy, and auditable configuration.
Review invites, directory sync, and provisioning as high-authority lifecycle operations with tenant, role, and approval checks.
Review prices, credits, refunds, and entitlements as server-owned state transitions that preserve an auditable ledger invariant.
Start an authorized hunt as a portable, local project with scope, rules, and evidence boundaries recorded before work begins.
Normalize saved recon into a provenance-preserving local inventory, then use it to expose coverage gaps without expanding authorization.
Turn an observation into a ranked, evidence-aware hypothesis with one clear test that could prove or reject it.
Run the smallest authorized check that resolves a claim, then retain an evidence-linked outcome even when the result is incomplete.
Stress-test a lead against benign explanations and missing proof before time or impact claims grow around it.
Build a concise finding from durable evidence while separating validated behavior, modeled impact, and unresolved uncertainty.
Package local hunt state so another researcher or client can resume from evidence, gaps, and ranked actions without repeating work.
No skills match that combination. Try a broader term or clear the filters.
Open a guide, switch to raw Markdown, or download its SKILL.md. Each file includes its review notes and source links. For cited research alongside it, use the MCP research tools.
| / | Focus search on this page |
| Ctrl K | Search everything (command palette) |
| 1-9 | Library, Generator, Playbooks, Gadgets, Checklists, Payloads, Bypasses, Utilities, Reports |
| 0 | AI / MCP connector |
| j / k | Move selection down / up |
| Enter | Expand / open selected |
| c | Copy primary command of selected |
| f | Toggle favorite on selected tool |
| Esc | Clear search / close |