Latest Agent Skills

Expands one confirmed or suspected vulnerability into a Trailmark graph neighborhood of variant candidates by finding sibling functions, shared callers and callees, common sensitive sinks, common entrypoint paths, interface implementations, override relationships, type/reference neighbors, and structurally similar nodes. Use after one issue is found to seed variant-analysis, semgrep-rule-creator, static-analysis, or manual review with graph-derived candidate locations.

Writes and reviews structured Lean 4 proofs and designs Lean libraries following Mathlib conventions. Use when proving theorems in Lean, formalizing mathematics or specifications in Lean 4, defining new types or definitions in a Lean library, reviewing Lean proofs for readability and maintainability, refactoring long tactic proofs into lemmas, filling in sorry placeholders in a Lean development, setting up CI or linters for a Lean project, diagnosing slow proofs or maxHeartbeats timeouts, or writing custom tactics, macros, or linters.

Performs graph-assisted triage of a single security finding, SARIF result, weAudit annotation, suspicious function, or report excerpt using Trailmark reachability, entrypoint paths, taint, privilege-boundary, blast-radius, caller/callee, and neighborhood evidence. Use when deciding whether one candidate issue is reachable, prioritizing a finding before PoC work, preparing evidence for exploit validation, or checking whether a static-analysis result is actionable.

Runs a Trailmark structural review gate over a branch, pull request, fix commit, release diff, or git ref range to detect new entrypoints, new tainted paths, removed validation or authorization calls, privilege-boundary drift, blast-radius growth, complexity growth, and newly reachable sensitive sinks. Use when reviewing a PR, branch, remediation commit, or release diff where graph-level security regressions should be checked before merge.

Selects bounded, graph-informed source slices with Trailmark and delegates focused code analysis or patch-proposal work to a smaller subagent. Use when offloading function-, class-, caller-, callee-, call-path-, entrypoint-, or line-focused code tasks to constrained or locally hosted models without exposing the full repository.

This skill should be used when the user asks to "open source this project", "prepare this repository for public release", "make this repo public", "check open-source readiness", "choose a license for this project", or "set up release automation" ahead of a public launch. Provides a release-readiness workflow covering secrets hygiene, licensing, documentation, CI, and language-specific packaging.

Triages a repository's open GitHub issues and pull requests via the gh CLI. Optionally reviews and merges ready PRs — incrementally merging passing automated/bot PRs and maintainer-approved ones, and spawning review subagents for never-reviewed ones — then closes already-resolved issues with comments citing the resolving PR or commit, cross-links issues with their pending fix PRs, and assigns local-only priority and change-size estimates for everything outstanding. Use when triaging, grooming, or reviewing a repository's open issues and PRs.

Runs an autonomous review-and-fix improvement loop over any code target — a skill, plugin, module, or directory — using a reviewer the user names: any installed skill or agent. Keeps a cross-round findings ledger, escalates when fixes stop converging, and guards scope mechanically. Use when asked to 'improve this code until review passes', 'run an improvement loop with <reviewer>', or to iterate review-and-fix with a specific reviewer. For skills prefer the skill-improver entry; for a branch prefer pr-improver.

Runs an autonomous review-and-fix improvement loop over the current branch's changes until a PR review comes back clean, scoped mechanically to the directories the branch touched. Reviews are performed by an installed PR-review skill (default: pr-review-toolkit's review-pr). Use to fix review findings on a branch before opening or updating a pull request ('clean up this branch', 'fix this PR until review passes', 'run review-and-fix on my changes'). NOT for a one-time review — run the PR-review skill directly.

Drafts copy-paste-ready /goal commands for goal mode in Claude Code and Codex. Use when the user asks to create, write, rewrite, improve, compress, clean up, or prepare a goal prompt, goal condition, /goal command, goal-mode objective, or copy-ready long-running task objective.

Guides C++ code toward modern idioms (C++20/23/26). Use when writing new C++ code, modernizing legacy patterns, or working on security-critical C++. Replaces raw pointers with smart pointers, SFINAE with concepts, printf with std::print, error codes with std::expected.

Behavioral guidelines to reduce common LLM coding mistakes. Use when writing, reviewing, or refactoring code to avoid overcomplication, make surgical changes, surface assumptions, and define verifiable success criteria.

Monitor competitor signals and surface strategic implications for your roadmap. Use when asked to monitor competitors, track the competitive landscape, produce a competitive briefing, or understand what has changed in the market this week or month. Produces a structured intelligence brief with high/medium/low priority signals, roadmap implications, and a strategic landscape summary. For a single competitor announcement use competitor-signal-tracker; for a one-off deep dive use competitor-teardown.

Present a design to a client so they can make a decision rather than react to a picture — the brief restated, the moves explained against it, the options with their trade-offs, and a specific decision requested. Use when asked to present a design to a client, prepare a design review meeting, present concept options, or when a client keeps changing their mind after sign-off. Produces the presentation structure, the brief-to-design traceability, the options with honest trade-offs, the cost and programme position, and the decision record that makes sign-off stick.

Decode a wedding vendor contract — venue, photographer, caterer, band — before signing: deposits and their refundability, cancellation and postponement terms, the substitute-performer and force-majeure clauses, and overtime math. Use when someone asks review this venue contract, is this photographer contract normal, what if we have to postpone, or decode this caterer agreement. Produces a clause decode with 🔴🟡🟢 severity, the cancellation-cost timeline, the questions to ask this vendor, and what's actually negotiable.

Draft answers to a vendor security questionnaire (SIG, CAIQ, or a custom sheet) from your real controls — fast, consistent, and honest about gaps. Use when asked to fill out a security questionnaire, answer a SIG/CAIQ, respond to a customer's security review, or complete a vendor risk assessment. Produces drafted answers grounded in your stated controls, a gap list of questions you can't truthfully answer yet, and reusable answer snippets for next time — never fabricated compliance.

Synthesises user signals from multiple research sources into a unified, weighted insight brief. Use when you have data from interviews, support tickets, NPS verbatims, app reviews, or sales calls and need to reconcile contradictions, surface the underlying need behind requests, or answer 'what are users really telling us'. Produces ranked insights with confidence ratings, source weighting rationale, divergent signal analysis by user segment, and a research gap identification section.

Decode a moving company estimate — binding vs non-binding, the weight and cubic-feet games, valuation vs insurance, and the red flags that precede hostage-load stories. Use when someone asks 'is this moving quote legit', 'decode my moving estimate', 'binding vs non-binding estimate', or 'how do I avoid moving scams'. Produces an estimate-type decode with what-you'll-actually-pay scenarios, ranked red flags, the valuation decode, and the questions that separate real movers from brokers.

Build a plan to keep a long-distance relationship close and healthy — communication rhythms, visits, shared experiences, and a shared sense of the finish line. Use when asked to help with a long-distance relationship, how to make LDR work, we're going long distance, or keep our relationship strong apart. Produces a communication rhythm that fits both schedules and time zones, ideas for shared experiences across the distance, a visit and cost plan, ways to handle the hard parts (jealousy, loneliness, resentment), and an honest 'the plan' conversation about the end goal.

Have the conversation about someone's end-of-life wishes — before a crisis forces it — gently, respectfully, and thoroughly enough to actually guide decisions later. Use when asked how do I talk about end-of-life wishes, discuss my parent's wishes, advance care planning conversation, or ask about their medical and final wishes. Produces a way to open this hard conversation without it feeling morbid or forced, the areas to cover (medical wishes, care preferences, where they want to be, what matters to them, practical/legal), how to listen rather than impose, and how to document and share what's decided — so their wishes are known and honored. Not legal or medical advice.

Verify work actually meets its brief BEFORE declaring it done — a structured self-review pass that catches the gaps, unmet requirements, and untested claims that 'looks finished' hides. Use before handing over any deliverable (document, code, analysis, plan), when past work kept coming back with 'you missed…', or as the standing final step of any multi-step task. Produces the verified deliverable plus a short verification record: what was checked, what was found and fixed, what remains open.

Write the design and access statement that supports a planning application — the site analysis, the design rationale, the policy compliance, and the answers to the objections the case officer will receive. Use when asked to write a design and access statement, planning statement, or supporting statement for a planning application, or to respond to a planning objection or refusal. Produces the site and context analysis, the design evolution narrative, the policy-by-policy compliance table, the access and amenity assessment, and the pre-empted objections. Not planning or legal advice; local policy and procedure govern.

Get real information out of a 15-minute parent-teacher conference — the questions that beat 'how's she doing', the data to bring from home, and the follow-up that makes the meeting matter. Use when asked prepare me for the parent teacher conference, what should I ask my kid's teacher, the conference is 15 minutes what do I prioritize, or how do I raise a concern without making it adversarial. Produces the prioritized question list, the home-observations brief, the concern-raising scripts, and the follow-up plan with owners.

Organise the document pile for a visa, work-permit, or residency application — what to gather, in what order, and the common rejection triggers to avoid. Use when asked to help with a visa application, build an immigration document checklist, prepare paperwork for a work permit or green card, or organise what an application needs. Produces the categorised document checklist, the gather-in-this-order plan, the common-mistake/rejection-trigger list, and the professional-help flags. Organises the paperwork; complements the visa-interview simulator.

Map a realistic path toward financial independence — the number you'd actually need, your savings rate's massive effect on the timeline, and the honest tradeoffs. Use when asked how do I reach financial independence, explain FIRE, what's my FI number, or plan for financial freedom. Produces an educational read on your rough FI number (and why the savings rate matters more than income), the timeline math at different savings rates, the levers and lifestyle tradeoffs, the different flavors (lean/coast/full), and the traps — turning a vague dream of freedom into a directional plan. Not financial advice.

Decide whether, when, how, and to whom to disclose a disability or health condition at work — weighing the real benefits (accommodations, protection, honesty) against the real risks (bias, gossip), tuned to your specific situation. Use when someone says 'should I tell work about my disability/condition', 'disclose my ADHD/chronic illness at work', 'when do I tell my employer', or 'how much do I share'. Produces a decision framework for the situation, a disclosure script if you choose to, and the minimum-disclosure options. Your choice throughout; it never pushes disclosure.

Evaluate performance fairly when output is AI-assisted — what still measures the human, what now measures the tooling, and how to run the review conversation. Use when reviewing someone whose work is heavily AI-assisted, when output volume stopped meaning anything, when calibrating a team with uneven AI adoption, or when writing review criteria for the AI era. Produces review guidance: a what-measures-whom analysis, rewritten criteria, calibration rules for mixed-adoption teams, and conversation scripts. For the general review document use performance-review; for redesigning the role itself use role-redesign-for-ai.

Generate the strategic story connecting a product roadmap to company goals in a form non-technical stakeholders can repeat. Use when asked to explain the roadmap, present strategy to leadership or the board, write the why behind the roadmap, create a narrative for all-hands, or make the roadmap tell a story. Produces a themed narrative with executive summary, progression arc, hard-question preparation, and what's-not-on-the-roadmap section.

Reread a harsh message, criticism, or perceived slight without the emotional spike — separate what was actually said from what your brain is amplifying. Use when asked this message really stung, am I overreacting to this, help me not spiral over this feedback, or did they mean it that way. Produces a calm read of what was literally said vs the story you've layered on, a check on the most likely (usually more neutral) intent, whether any action is actually warranted, and a grounded response option — easing the disproportionate sting that rejection-sensitive brains feel.

Audit and optimize a marketplace listing (Amazon, Etsy, eBay, Walmart) to rank and convert. Use when asked to optimize an Amazon/Etsy listing, improve marketplace SEO, fix a product listing that isn't selling, or write keyword-rich titles and bullets. Produces a prioritised optimization — title, bullets, backend keywords, A+/description, images plan, and conversion fixes — mapped to how that marketplace ranks and shoppers decide.

Decode a home renovation or repair quote — allowances that aren't prices, exclusions that become change orders, payment schedules that shift risk, and what a comparable-bids check should cover. Use when someone asks 'is this contractor quote fair', 'decode this renovation bid', 'what should be in a contractor contract', or 'why do these three bids differ so much'. Produces a section-by-section decode, the allowance and exclusion audit, payment-schedule risk analysis, and the questions that make bids comparable.

Explain a financial statement (P&L, balance sheet, or cash flow) in plain English. Use when asked to explain a P&L / income statement, a balance sheet, a cash flow statement, or to make financials understandable to a non-finance reader. Produces a plain-language walkthrough — what each section means, the line items that matter, the key ratios, and the story the numbers tell — so a non-accountant can read and act on it. Not financial advice.

Document a pharmacist's clinical intervention — the interaction or error spotted, what was done, what the prescriber decided, and the outcome — so the record shows the catch and the care that followed. Use when asked to document a clinical intervention, record a prescriber call about an interaction, log a near-miss, or build an intervention record for audit or remuneration. Produces the intervention record with severity and evidence, the prescriber-call script, the outcome and follow-up, and the aggregate reporting fields. A documentation framework for a licensed pharmacist; every interaction and severity judgement must be verified against current references.

Investigate a commercial vehicle incident so the record is defensible and the cause is actually found — evidence secured in the first hours, the preventability decision made on stated criteria, and corrective action aimed at the system rather than the driver. Use when asked to investigate a vehicle accident, run a post-accident review, decide whether a collision was preventable, or write an incident report for a fleet. Produces the evidence checklist, the sequence reconstruction, the preventability determination with reasoning, the contributing-factor analysis, and the corrective actions. Not legal advice; involve your insurer and counsel early.

Build a community management playbook for a brand's social media channels. Use when asked to create guidelines for managing comments, DMs, and community interactions, define a moderation policy, or build response frameworks for social media community managers. Produces a complete playbook with response templates, escalation paths, moderation rules, and tone guidelines.

Turn a client's wishes into an architectural brief that can actually be designed against — the accommodation schedule, the adjacencies, the budget reality, and the constraints named before they become change orders. Use when asked to write an architectural brief, prepare a project brief for a building, run a client briefing, or capture requirements for a new build or refurbishment. Produces the accommodation schedule with areas, the adjacency and circulation requirements, the site and statutory constraints, the budget and programme reality check, and the explicit exclusions.

Build a maintenance schedule for your car so it stays reliable and holds value — the service intervals, the DIY-vs-shop split, and the checks that prevent breakdowns and rip-offs. Use when asked for a car maintenance schedule, what maintenance does my car need, how to keep my car running, or am I being upsold at the mechanic. Produces an interval-based schedule keyed to your vehicle and driving, the essential do-not-skip items, DIY vs professional, seasonal checks, and how to spot unnecessary upsells — flagging that your owner's manual is the authority.

See an idea through two committed extremes — a true believer and a hard skeptic — so you get the full range before settling in the middle. Use when asked should I believe this, is this hype or real, give me both sides, or how excited should I be about. Produces the believer's fullest bull case and the skeptic's sharpest bear case (each committed, not hedged), the crux question that separates them, and a grounded read on where the truth probably sits — great for evaluating claims, trends, opportunities, and your own enthusiasm.

Map stakeholders for a product decision and produce a tailored influence strategy with talking points. Use when asked to get alignment, build consensus, get buy-in from engineering or finance or legal, navigate organisational resistance, or plan stakeholder conversations for a major initiative. Produces a stakeholder map, recommended conversation sequence, and tailored talking points per stakeholder.

Analyze a rental / investment property's returns — cash flow, cap rate, cash-on-cash, ROI. Use when asked to analyze a rental property, evaluate a real-estate investment, run the numbers on an investment property, or compute cap rate / cash-on-cash. Produces an investment analysis — income and expenses, NOI, cap rate, monthly cash flow, cash-on-cash return, and a verdict against the investor's criteria — with formulas and a worked example. Not financial advice.

Handle a prescription you are not going to dispense — the professional judgement recorded, the patient conversation, the prescriber contact, and the continuity of care that stops a refusal becoming abandonment. Use when asked how to refuse to fill a prescription, decline to dispense, handle a suspected forgery or early refill, or document a corresponding-responsibility decision. Produces the decision record, the conversation script, the prescriber contact, the continuity step, and the escalation path. Legal obligations vary by jurisdiction; verify yours.

Cut through passive-income hype to what's actually realistic for you — the real effort, capital, and risk behind each option, and which (if any) fit your situation. Use when asked how do I make passive income, is passive income real, best passive income ideas, or help me build income streams. Produces an honest teardown of the popular passive-income options (what they really require, how 'passive' they actually are, typical returns and risks), a match to your capital/skills/time, the scams and get-rich-quick traps to avoid, and a grounded next step — replacing the fantasy with a realistic path. Not financial advice.

Set up a system to manage medications safely — for yourself or someone you care for — so doses aren't missed, doubled, or dangerously combined. Use when asked help me manage medications, keep track of my parent's pills, set up a medication system, or I keep forgetting my meds. Produces an organized medication list (what, dose, when, why), a routine and reminder setup that fits the person, a refill-tracking method so nothing runs out, safety checks (interactions and duplications to raise with a pharmacist), and an emergency-ready summary — because medication errors are common and dangerous, and a system prevents most of them. Not medical advice.

Prepare to get the most out of a medical appointment — for yourself or someone you care for — with the right questions, the information to bring, and how to make sure you're heard. Use when asked help me prepare for a doctor's appointment, questions to ask the doctor, advocate for my parent at the doctor, or how do I get the most from this appointment. Produces a focused list of what to raise and ask (prioritized, since time is short), the history and info to bring, note-taking and 'teach-back' tactics so you actually understand, how to speak up if dismissed, and what to confirm before leaving — not medical advice, but better navigation of care.

Prepare for an informational interview — the outreach, the questions, and the follow-up — so a 20-minute chat actually helps your career instead of wasting their time. Use when asked to prep for an informational interview, questions to ask someone in [field], how to reach out for a career chat, or coffee chat prep. Produces a low-friction outreach message, a focused question set tuned to your goal (exploring a field, breaking in, a specific company), how to run the conversation, what NOT to do (don't ask for a job), and a follow-up that keeps the relationship warm.

Decode a disability insurance policy or employer LTD plan — own-occupation vs any-occupation, the benefit math after offsets and taxes, and the definitions that decide whether it pays. Use when someone asks 'is my disability insurance any good', 'decode my LTD policy', 'what does own-occupation mean', or 'how much would I actually get'. Produces a definition decode of the clauses that decide claims, the real benefit math after offsets, ranked red flags, and the questions to ask before relying on the coverage.

Resolve a dependency or version conflict (npm, pip, yarn, pnpm, Maven, Go modules) step by step. Use when an install fails with peer-dependency or version-conflict errors, packages won't co-exist, or a lockfile is fighting you. Produces the conflict explained, the resolution options ranked by safety, exact commands, and how to keep it from recurring.

Audit a landing page or funnel step and produce a prioritised CRO test plan. Use when asked to improve conversion rate, audit a landing/signup/checkout page, reduce funnel drop-off, or plan A/B tests for a page. Produces a CRO plan — a heuristic conversion audit, the diagnosed friction, prioritised test hypotheses (ICE), test designs with sample-size math, and the measurement guardrails.

Write cold outreach to potential clients that gets replies — specific, useful, and about them — instead of the templated pitch that gets deleted. Use when asked to write a cold email to a prospect, get clients through outreach, cold pitch help, or reach out to potential customers. Produces a researched, personalized message that leads with their problem, a clear low-friction ask, proof you're credible without bragging, a subject line, and a short follow-up sequence — plus who to target and what to avoid so it lands as a helpful note, not spam.

Run a family meeting to make a big care decision together — so it's a shared, informed decision instead of a fight or one person deciding alone. Use when asked help me run a family meeting about care, we need to decide together about mom's care, my siblings and I disagree about our parent, or facilitate a care decision. Produces an agenda and structure for the meeting, how to prepare (facts, options, everyone's input), ground rules to keep old family dynamics from derailing it, how to include the person being cared for, ways to work through disagreement toward a decision, and clear next steps with owners — turning an emotional, conflict-prone conversation into a productive one. Not legal, medical, or financial advice.

Respond to a brand or executive impersonation incident — deepfaked executives, cloned support lines, fake apps, spoofed domains, or AI-generated scam content wearing your name. Use when a deepfake of a leader is circulating, customers report a fake version of your product or support channel, or to prepare the impersonation playbook before it happens. Produces an incident response: verification protocol, takedown sequencing by platform, customer and public communications, and the hardening plan. For general crisis comms use press-release/pm-crisis skills; for security incidents inside your systems use security-incident-response.

Decode an auto repair estimate — what each line actually is, which items are urgent vs upsell, and the questions that separate a fair shop from a fishing expedition. Use when someone asks 'is this repair quote fair', 'decode my mechanic's estimate', 'do I really need all this', or 'is the shop ripping me off'. Produces a line-by-line decode with urgency triage, parts/labor sanity checks, ranked red flags, and the exact questions to ask the shop before authorizing.

Create an Architecture Decision Record (ADR) for any technical decision. Use when asked to document a technical decision, write an ADR, record an architecture choice, or capture why a technology or approach was selected. Produces a structured ADR with context, decision, consequences, and tradeoffs.

Design and facilitate any workshop, working session, or collaborative meeting. Use when asked to plan a workshop, design a facilitated session, run a ideation session, or create a workshop agenda. Produces a complete facilitation guide with session design, activity instructions, timing, and materials.

Present a dental treatment plan the patient actually understands and accepts — sequenced by clinical priority, costed with their coverage applied, and explained in the order that answers what-is-wrong before what-it-costs. Use when asked to present a treatment plan, improve case acceptance, explain treatment to a patient, or handle a patient who says they want to think about it. Produces the phased plan, the plain-language explanation per phase, the cost and coverage breakdown, the consequences-of-delay framing, and responses to the four common objections. Clinical decisions remain the treating clinician's; this shapes the conversation, not the diagnosis.

Decode a severance agreement before you sign it — what you're giving up, what's negotiable, and the deadlines that decide your leverage. Use when asked to decode my severance, is this severance offer normal, review my separation agreement, or should I sign this release. Produces a clause-by-clause decode with ranked red flags, the money math (severance vs what you're releasing), the consideration-period clock, and the asks worth making.

Set criteria-based return-to-activity milestones so the decision rests on what the patient can demonstrate rather than on how long it has been — the test battery, the thresholds, the graded reintroduction, and the honest conversation about residual risk. Use when asked to plan a return to sport or work, set return-to-play criteria, decide whether someone is ready to return, or handle pressure to clear someone early. Produces the criteria by stage, the test battery with thresholds, the graded exposure plan, the residual-risk discussion, and the decision record. The clearance decision is the treating clinician's, made with the patient and the relevant medical team.

Turn a renovation idea into a realistic scope, budget, and sequence before you hire anyone — so you go in informed instead of getting sticker-shocked or scoped. Use when asked to plan a renovation, budget for a remodel, how much will renovating [X] cost, or scope my home project. Produces a scoped breakdown of the work, a realistic budget range with a contingency, the sequence and rough timeline, must-decide-early choices, where costs balloon, and what to line up before getting quotes — flagging that local prices and permits vary, so verify with real quotes.

Announce a price increase without triggering a churn spike — the rationale, grandfathering, effective dates, the FAQ, and the internal brief so support isn't blindsided. Use when asked to announce a price increase, write pricing change comms, raise prices to customers, or communicate new pricing. Produces the customer email, the FAQ with objection handling, the grandfathering/transition terms, and the internal enablement brief. A pricing-comms craft, not a discount.

Work down the cost of a prescription systematically — the generic and therapeutic-alternative conversation, discount programs vs insurance math, pharmacy price variance, and manufacturer/assistance programs, in the order that saves the most first. Use when asked my prescription is too expensive, how do I save on my meds, is there a cheaper version of this drug, or I can't afford my medication. Produces the cost-reduction ladder for the specific prescription, the scripts for pharmacist and prescriber conversations, and the never-do list (skipping doses is not a savings plan).