Security

unboundcompute Skill

Audit an over-the-air or firmware update channel for a device that accepts an image it should reject: an update whose signature is not verified so an attacker installs arbitrary firmware, an update fetched over an unauthenticated transport an on-path attacker can swap, a rollback to an older signed image with known vulnerabilities because the device does not enforce version monotonicity, an update server or manifest URL the device trusts without authentication, and an unencrypted image that leaks secrets and eases reverse engineering. Covers IoT and embedded devices, routers, wearables, and any product that fetches and installs firmware or application updates in the field. Use when a device installs firmware it fetches and the verification of that image before it runs is the boundary. The unsigned, swapped, or rolled-back update is the source, the persistent code execution on the device is the sink, and the missing signature check, transport authentication, or rollback protection is the bug.

unboundcompute Skill

Audit payment provider callbacks and settlement notifications for the trust that lets an attacker forge or alter a payment result: a callback whose signature is not verified so a spoofed success is accepted, an amount or currency taken from the callback or client rather than reconciled against the order the server created, a success notification not bound to a specific order so it can be replayed onto another order, and a settled status trusted without confirming it out of band with the provider. Covers redirect returns, server-to-server webhooks, and status polls where a payment processor tells the application a charge succeeded. Use when the application learns a payment result from an external processor and that message gates fulfillment. The forged or altered payment notification is the source, the order marked paid and fulfilled is the sink, and the unverified signature, unreconciled amount, or unbound order reference is the bug.

unboundcompute Skill

Audit payment and checkout state machines for transitions an attacker can drive out of order or replay for value: an order marked paid before the charge is confirmed, a step that can be skipped or repeated so goods ship without settlement, a non-idempotent charge or fulfillment endpoint that double-processes on a retried or replayed request, and a refund or cancel that returns value while the underlying charge stays captured. Covers checkout, charge, fulfillment, and refund flows where money and goods change hands across a sequence of state transitions. Use when a purchase moves through ordered payment states and the transitions and their idempotency are the boundary. The out-of-order or replayed transition is the source, the value released without settlement is the sink, and the skippable step or non-idempotent handler that allows it is the bug.

unboundcompute Skill

Audit an ERC-4337 account-abstraction deployment for trust misplaced in the user-operation lifecycle: a smart account whose validation accepts a signature or nonce it should reject, a paymaster that agrees to sponsor gas for operations it should not so an attacker drains its deposit, a bundler or entry-point assumption that lets a user operation be replayed or reordered for gain, and validation logic that reads mutable state or reaches outside its allowed scope. Covers smart-contract wallets, paymasters, bundlers, and the entry point in an account-abstraction stack where a user operation is validated and sponsored before it executes. Use when a user operation is validated, paid for, and executed by separate parties and that trust split is the boundary. The crafted user operation is the source, the drained paymaster or unauthorized execution is the sink, and the over-permissive validation or sponsorship rule is the bug.

unboundcompute Skill

Audit error handling and diagnostic surfaces for sensitive information a real client receives, where an exception path, a debug feature, or a diagnostic endpoint returns stack traces, database errors, internal paths, framework or version banners, configuration, secrets, or messages that differ enough to enumerate users. Use when reviewing how a service responds to malformed, unauthorized, or failing requests in its deployed configuration, and whether debug modes, source maps, or version-control metadata are exposed. Scoped to what production actually returns, not developer-only verbosity. The error or diagnostic path is the source, the response, header, or user-visible log is the sink, and disclosing detail that aids a further attack is the bug.

unboundcompute Skill

Audit both directions of webhook trust: an inbound handler that acts on a payload without proving it authentic, and an outbound fetch of a caller-supplied URL that reaches internal targets. Covers inbound handlers with no signature check, a signature compared in non-constant time, a signature computed over a re-serialized body instead of the exact raw bytes, a verification result that is computed but never enforced, and no timestamp or replay defense; and outbound callback or fetch URLs validated by substring or blocklist, or by a single pre-connect lookup that a redirect or a rebind defeats. Use when reviewing code that receives a signed webhook and performs a state change, or that fetches a URL the caller controls. The inbound request or the caller-supplied URL is the source, the state-changing handler or the server-side fetch is the sink, and a missing or bypassable trust check between them is the bug.

unboundcompute Skill

Hunt injection into database connection strings and JDBC or driver URLs where untrusted input sets the host, a driver property, or a URL parameter, turning a data connection into a request to an attacker server or an unsafe driver feature. Covers a tenant, hostname, or option taken from input and spliced into a connection URL, driver properties that enable local file reads, arbitrary command execution, or class loading, and multi-attribute connection strings where an extra property overrides a security setting. Use when an application builds a database or service connection string from user or tenant input rather than from fixed configuration. The untrusted value that becomes a connection host or property is the source, the connect call is the sink, and the dangerous driver feature or redirected endpoint it reaches is the bug.

unboundcompute Skill

Hunt machine credentials that are live, over-privileged, and actually reachable, not just present. Covers non-human identities and secrets across code, configuration, and infrastructure definitions: API keys, service-account credentials, and long-lived tokens. Separates a secret that merely exists from one an attacker can reach and use, and adjudicates each by whether it is still valid, how much it grants, and whether an untrusted path leads to it. Use when reviewing secret exposure, machine identities, or the blast radius of a leaked credential, and when a scanner reports many secrets and you need to know which ones matter. A reachable, live, over-privileged credential is the finding.

unboundcompute Skill

Audit whether a mobile app's integrity and environment checks actually enforce a security decision, where the app runs a root or jailbreak check, a repackaging or debugger check, or an emulator check but decides the result on the client, so a controlled device defeats it, or the app relies on such a check to protect a secret or an operation that a genuine attacker on their own device simply bypasses, because the check is advisory rather than a real control and the protected asset is reachable once it is defeated. Use when a mobile app performs environment or tamper checks and treats their outcome as a security boundary. Covers client-decided integrity checks, checks not bound to a server-verified attestation, and secrets or operations gated only by such a check. The controlled or tampered device the check cannot truly detect is the source, the client integrity decision is the sink, and reaching the protected asset by defeating the check is the bug.

unboundcompute Skill

Hunt security bypasses where a check passes on one representation of untrusted input and a normalization, decoding, or case-folding step then turns it into a different, dangerous form at the sink. Use when an authorization, filter, allowlist, path-containment, or identifier comparison runs on input that is later normalized, percent- or entity-decoded, or case-folded before it reaches the operation it guards. Covers Unicode normalization forms, overlong and double encoding, case-fold and locale collisions, and mixed-script or homoglyph identifiers. The input that passes the check in one form is the source, the security decision made on the wrong form is the sink, and the bypass or identity confusion that the later transform enables is the bug.

quantumquirkxyz Skill

Analyse DeFi protocols — AMMs, lending, stablecoins, derivatives — across protocol mechanics, economic security, and composability risks.

quantumquirkxyz Skill

Run a backtest with full audit hygiene — biases, costs, out-of-sample, regime splits — and produce a verdict on whether a strategy is robust.

quantumquirkxyz Skill

Analyse and compare consensus mechanisms — PoW, PoS, BFT, DA — with security properties, finality guarantees, and economic incentive alignment.

quantumquirkxyz Skill

Design and audit on-chain and off-chain governance — token voting, delegation, quadratic voting, timelock, multisig, and attack vectors.

inbharatai Skill

Detect hardcoded secrets and credentials in codebases — API keys, tokens, passwords, and remediation steps.

inbharatai Skill

Perform threat modeling — STRIDE, attack trees, data flow diagrams, and mitigation recommendations.

inbharatai Skill

Write information security policies — acceptable-use, incident response, data classification, and access control.

inbharatai Skill

Audit code for security vulnerabilities — OWASP Top 10, injection flaws, auth issues, secrets exposure, and dependency risks.

inbharatai Skill

Audit project dependencies — outdated packages, known CVEs, license conflicts, and upgrade recommendations.

inbharatai Skill

Write optimized Dockerfiles — multi-stage builds, layer caching, security hardening, and minimal image sizes.

m2ai-portfolio Skill

Audit your Claude Code skills directory, CLAUDE.md, and recent session patterns to surface candidates for new skills -- gaps, redundancies, and formalization opportunities -- and to catch skill hygiene failures: runtime artifacts (venv, __pycache__, node_modules) living inside a skill directory, and absolute paths or identity leaking out of tracked files.

m2ai-portfolio Skill

Run Nate's three-question industry diagnostic — what inefficiency is this built on, how fast can AI close it, what new gap does closure create — and output a migration map showing where value moves upstream. Use when the user wants a fast (5-minute) diagnostic of any industry, product, or workflow without running the full arbitrage-audit.

m2ai-portfolio Bundle

Three-phase audit to determine whether a software tool or artifact should be demoted down the production-class ladder (Customer-Facing → Supported Internal → Team Beta → Personal Tool → retired). Returns a Demote/Hold verdict with explicit trigger evidence, transition requirements, and a "maintenance cost no one is watching" estimate. Trigger on "demotion audit", "should I demote this tool", "is this tool still worth supporting", "downgrade this tool", "retire this tool".

m2ai-portfolio Skill

Scan a codebase to map where human comprehension has gone missing -- identifying modules with no decision logs, no behavioral contracts, no authorship trail, and high AI-generation signals. Produces a tiered darkness report (summary heat map + ranked TODO list) scoring each module 0-100. Use when onboarding to a new repo, auditing AI-generated code risk, assessing technical debt from comprehension gaps, or as a periodic health check. Trigger on "dark code", "darkness audit", "comprehension map", "who understands this code", "AI-generated code risk", "code understanding gap".

m2ai-portfolio Bundle

Audit Claude Code skills for content quality against Anthropic best practices. Use when Forge runs maintenance cycles, when the user asks to check skill quality, or when reviewing skills before publishing.

m2ai-portfolio Bundle

Accumulates repo-specific code review lessons and surfaces them before the next review of the same file or module. Learns from recurring issues -- migration bugs, error-handling patterns, fixture quirks, security gaps -- so the same mistake isn't flagged for the first time twice. Trigger phrases: "review this", "check this PR", "code review", "review these changes", "what do we know about this file".

m2ai-portfolio Bundle

Set up and configure Claude Code remote access via Telegram and Discord channels. Guides through bot creation, plugin installation, security lockdown, and always-on configuration.

m2ai-portfolio Bundle

Score an AI deployment against six implementation layers (workflow, data, authority, evaluation, audit trail, recovery). Outputs a tier verdict and surfaces the two questions your next enterprise buyer will ask that you cannot yet answer. Use on "impl-arch-audit", "implementation architecture audit", "score my AI deployment", "six-layer audit".

m2ai-portfolio Bundle

Audit the next N days of a Google Calendar and surface hygiene issues before the user notices them -- back-to-back meetings with no buffer, missing prep blocks before important events, overloaded days, and overlapping commitments. Proposes specific calendar edits. Use when asking "calendar hygiene", "audit my calendar", "check my schedule", "what's wrong with my week?", "add buffers to my calendar", "prep blocks", or wanting anticipatory calendar management rather than reactive scheduling.

mariourquia Bundle

Executive finance leader. Consumes reporting_finance_ops_lead rollups, portfolio_manager rollups, and asset_manager covenant views. Owns treasury posture, covenant-cushion decisions, refi / recap / workout decisions (jointly with CEO and portfolio_manager), fund-accounting policy, investor- and lender-facing finals sign-off (jointly with CEO where required), and audit / fund-document actions.

mariourquia Bundle

The plumbing for every gated action. Takes a pending `ApprovalRequest` and routes it to the correct approvers per the approval matrix row and org overlay, tracks SLA, and records the outcome to the approval audit log. Invoked by any workflow that opens an approval request. Never decides on behalf of humans; composes and routes.

mariourquia Bundle

Coordinates every step from notice-to-vacate through unit transfer to the turn workflow. Produces the move-out inspection checklist, assembles the final ledger and statutorily-compliant security-deposit statement, handles damage charge disputes, and hands off to the turn workflow. Enforces jurisdiction-specific return deadlines.

mariourquia Bundle

Owner-oversight review of third-party property manager performance. Scores against the PMA's service standards and the overlay's rubric, computes adherence metrics, flags audit findings, produces the scorecard memo and the TPM-facing communication draft. Runs monthly per property and quarterly per PMA. Material performance gaps and any PMA amendment or termination are gated.

mariourquia Bundle

Evidence-driven implementation intake workflow for residential multifamily and related CRE operating environments. Guides sponsors, implementation leads, data teams, operations leaders, and reporting owners through source inventory, access readiness, export evidence, field-level semantics, crosswalk ownership, reporting calendar, approvals, blockers, assumptions, and final leader sign-off. Produces implementation packets that are usable by both executives and build teams, while refusing silent assumption upgrades and refusing secret storage.

evolution-foundation Bundle

Query PostgreSQL databases configured in .env (DB_POSTGRES_N_*). Use when the user asks to query, explore, or audit data in a Postgres database. Picks connection by label (e.g. 'msgops-dev', 'bms-prod') or numeric index. Read-only by default — writes refused unless DB_POSTGRES_N_ALLOW_WRITE=true on that block.

evolution-foundation Skill

Send email messages with explicit user confirmation. Requires the exact confirmation string "YES, SEND" before executing send operations. Logs all sends to audit trail. Use ONLY when user explicitly confirms they want to send an email, never auto-invoke. This is a critical write operation requiring human approval.