More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Blindspot Pass: Pre-Work Territory Briefing

    1

    Before you touch unfamiliar code or a new domain, get the map: the landmines, the inherited constraints, what good looks like here, and a sharper request — in thirty minutes.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Blindspot Pass: Pre-Work Territory Briefing

    Blindspot Pass: Pre-Work Territory Briefing

    Example session with this skill installed

    I need to add rate limiting to our webhook receiver. I've never worked in this part of the codebase. Brief me before I touch anything.

    • Read your context and instructions
    • Compiled the blindspot pass

    DEPTH CALIBRATION: Familiarity 1 (knows the domain, not this codebase) + blast radius 1 (module-level) = full five-section briefing.

    WHERE YOU WILL STEP WRONG: webhooks/ has TWO receivers: v1 (legacy, still gets 30% of traffic) and v2. Limiting only v2 leaves the hole open. The repo's limiter middleware reads the forwarded-for header; webhooks arrive through the internal proxy, so every sender looks like the same internal address.

    WHAT IS ALREADY DECIDED: Webhook handlers must acknowledge within 2 seconds (the sender retries after that). Any limiting must fail fast, never queue.

    WHAT GOOD LOOKS LIKE: The token-bucket helper in the internal API directory — per-key, with the sender-ID header pattern to copy.

    EXPERT QUESTIONS: 1) Per-sender or global limit? Best guess: per-sender; the abuse case is one noisy customer. Confirm with ops. 2) What happens on a 429 — does the sender back off? Unknown; check the provider's docs before choosing the status code. 3) Does v1 share the same ack budget as v2? Best guess: yes. Confirm from the receiver config.

    SHARPER REQUEST: "Add per-sender token-bucket rate limiting to BOTH webhook receivers (v1 and v2), keyed on the sender-ID header, failing fast with a 429 inside the 2-second ack budget. Mirror the internal API throttle helper." The original request named one receiver and no constraints; the rewritten one names both, the key, the failure mode, the time budget, and the reference implementation.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    The most expensive mistakes in unfamiliar territory are not typos — they are confident actions based on a wrong map. This skill produces the map first: a depth calibration that scales the briefing to the goal's blast radius, a real walk of the territory (code, commit history, local conventions, prior art, the domain's table stakes), and a five-section briefing that ends at understanding.

    Section by section: where you will step wrong (named potholes with file pointers, not generic advice), what is already decided (invariants and recorded decisions you inherit), what good looks like here (concrete examples from this repo), the questions an expert would ask (with labeled best guesses), and the sharper request — the before and after, where the gap is the value. A quality rubric scores each section before handoff, and the gate is strict: every claim carries a pointer, guesses stay labeled, nothing gets implemented inside the briefing, and thirty minutes of looking is the limit. "No significant blindspots" is a valid result.

    • A briefing is only as good as the territory access — without the code or docs, it degrades to labeled guesses.
    • It ends at understanding; it does not implement, estimate, or approve the work.
    • Thirty minutes of looking is a hard cap, so deep architectural unknowns may surface as expert questions rather than answers.
    • If the named expert is unavailable, guesses stay guesses — the briefing says how to verify them, it does not verify them for you.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 3 days ago

    Permissions

    No special permissions declared or detected

    Listed3 days ago

    What's inside

    Frequently Asked Questions

    Reading is the raw material; the briefing is the product. It forces the output into five sections with pointers, orders landmines by blast radius, and ends with a rewritten request — the difference between having looked and having a map someone else can use.

    Then the briefing says so plainly and you move on. A clean bill of health is useful information, not a failed exercise — and inventing landmines to justify the pass is an anti-pattern the skill names.

    A briefing that takes longer than the task would is theater. The cap forces the scout to prioritize by blast radius and to write down what is unknown instead of researching it all.