More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Self-Review Evaluator: Rate Against Your Career Ladder

    1

    Asked to assess a self-review, an AI reads it generously and hands back a flattering rating that mis-levels the person. Your agent starts every dimension at Not Evidenced and only raises it on a quote.

    Free

    Security scanned
    Self-Review Evaluator: Rate Against Your Career Ladder

    Self-Review Evaluator: Rate Against Your Career Ladder

    Example session with this skill installed

    Evaluate my self-review against our engineering ladder. I'm a Senior Engineer (L3). Ladder and review pasted below.

    • Read your context and instructions
    • Compiled the self-review evaluator

    The agent reads the ladder, identifies L3 and L4, and maps three projects from the review. Most cells start Not Evidenced; two move to At Level on quotes about owning a migration plan. It asks 9 probing questions about scope, independence, impact, and visibility. After your answers it upgrades two more cells and shows the quotes behind each change. The assessment quotes the L3 and L4 language for each dimension and marks Influence as Not Met for L4, with the gap in the ladder's words. It then gives the 30-day and 60 to 90 day plans and four rewording suggestions.

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

    About this skill

    About this skill

    An unguarded agent reading a self-review infers scope and impact that were never stated, gives credit for work that is only implied, and returns an assessment that feels good and is not useful. If the person relies on it for a promotion case, the gaps show up in calibration instead.

    This skill builds the guard into the structure. You supply your company's leveling framework (a file, pasted text, or a URL), your role and level in the framework's own names, and the self-review. Every dimension starts at Not Evidenced, and a rating moves up only on a specific quote from the review or from your answer to a follow-up question. The agent maps each project to the framework's own dimensions with quoted evidence, asks 2 to 4 targeted questions for each thin spot, then quotes the framework's current-level and next-level language per dimension and states Met, Partially Met, or Not Met with the specific gap. It ends with a 30-day plan to solidify the current level, a 60 to 90 day plan for the 2 or 3 biggest gaps, and 3 to 5 rewording suggestions for work you actually described.

    The proof

    There is no measurement for this skill, and the listing claims none. Its checks are rules you can audit in the output: every rating cell carries a quote, or it reads "Not Evidenced: needs probing"; every gap statement uses the framework's exact words; and rewrite suggestions never add work the review did not describe. If the framework cannot be read, the skill stops instead of rating from general knowledge. It pauses for you after inputs, after the probing questions, and before the plan.

    What you get

    SKILL.md with the calibration rules, rating scale, four-step flow, and gotchas, and references/example-ladder.md, a short generic five-level individual-contributor ladder (scope, impact, craft, execution, influence) written for this skill and not taken from any company. Results against it are labeled as generic.

    How to use

    1. Add the skill folder to your agent's skills directory (for Claude Code, ~/.claude/skills/self-review-evaluator/).
    2. Say "evaluate my self-review" and give it your leveling framework, your role and level, and the review text.
    3. Answer the probing questions, check what changed in the re-mapped ratings, then read the gaps and the plan.

    Source and license

    Free to use and MIT licensed. The full source, install steps, and issue tracker are on GitHub at https://github.com/robertmnyborg/self-review-evaluator.

    Ratings are only as meaningful as the framework you supply; results against the example ladder are generic. It rates only what is written in the review and your answers. It does not replace a manager's or calibration committee's judgment. If the framework cannot be read, it stops. No measured results yet.

    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

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 8 days ago

    • Free with an Agensi account

    Requires

    Any agent that loads SKILL.md files
    Your company's leveling framework as a file, pasted text, or readable URL (optional; a generic example ladder is included)

    Permissions

    File Scopes

    the
    framework
    and
    review
    files
    you
    point
    to
    (read)

    Reads only what you provide.

    Listed8 days ago

    What's inside

    Frequently Asked Questions

    An agent will assess a review if asked, and it tends to be generous. This skill defaults every dimension to Not Evidenced and only raises a rating on a quote, so the result reflects what the review shows.

    It offers the included generic five-level ladder and labels every result "rated against a generic example ladder, not your company's framework".

    Yes. It is markdown only, with no scripts, so any agent that reads SKILL.md can use it.

    No. Rewrite suggestions only reframe work you described in the review or in your answers.

    On GitHub at https://github.com/robertmnyborg/self-review-evaluator, under the MIT license. Every file in this package is there, plus the README and install steps.

    New in Business & Operations