agentleFS
leoleolroleoleolro/career-os/.claude/skills/decide/SKILL.md

decide

Skill0 starsChanged 55 days agoMITView on GitHub

Research and reason through a career decision (should I take this path, switch lanes, stay or go, accept this offer vs that one) with live evidence, source grading, a 20-year view, and kill criteria. Use when the candidate asks "should I", "which is better", "what would you do", "is this a good jump", "deep research on my career", "/decide".

56 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use redesign, experiment and 8 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • arjunkshahSkill

    redesign

    Redesign existing interfaces to premium quality while preserving product goals and functional behavior.

    Also in 3 projects that use this

  • SethGammonSkill

    experiment

    Automated optimization loop with scalar fitness function. Proposes changes in isolated worktrees, measures with a metric command, keeps improvements, discards failures. Supports convergence detection and diminishing returns.

    Also in 3 projects that use this

  • AnthropicSkill

    brief

    Generate contextual briefings for legal work — daily summary, topic research, or incident response. Use when starting your day and need a scan of legal-relevant items across email, calendar, and contracts, when researching a specific legal question across internal sources, or when a developing situation (data breach, litigation threat, regulatory inquiry) needs rapid context.

    Also in 5 projects that use this

  • genomewalkerSkill

    health

    Soul system health check with remediation

    Also in 3 projects that use this

  • AWSSkill

    map

    Map verified service capabilities against customer-provided control framework and threat model. Produces mapping-results.json, controls-catalog.md, and framework-mapping.md. Supports pluggable frameworks: CCMv4, NIST 800-53, CIS Benchmarks, ISO 27001, etc. Split into 4 focused sub-skills for better accuracy; this skill orchestrates them.

    Also in 3 projects that use this

  • HeshamFSSkill

    loop

    Run a command or check on a recurring interval

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 55 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Run 6 to 10 targeted WebSearch queries covering:

    line 14
  2. Very specific:

    Run node tools/signals.mjs for the candidate's own market data: what the hunt has actually seen, which beats any national statistic for …

    line 21
  3. Distinctive:

    Never invent a statistic or a salary band.

    line 40
  4. Distinctive:

    Never let a recruiter's number stand unqualified.

    line 41

“Common” is measured: kept by at least 6 in 10 of the 55 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

/decide <the decision>

Produces reports/career-decision-<today>.md and appends the verdict to profile/decisions.md. This is the agent's heaviest thinking task: do it properly or say why you can't.

0. Frame the decision first (do not skip)

Write down, before any research: the options as they actually exist (with links to live postings where they are jobs), the horizon the candidate cares about, the criteria that matter to him (from profile/career-context.md and preferences.json → priority_model), and what he has already decided that this must respect (for example the years-as-preference rule, stay/go factors, confidentiality). If an option is imaginary ("maybe an AI startup"), say so and treat it separately from live options.

1. Research, in parallel, with numbers

Run 6 to 10 targeted WebSearch queries covering:

  • Demand for each path (posting volumes, growth rates, talent shortage ratios; prefer analyses of large posting datasets).
  • The industry's own health (revenue, headcount, layoffs, government or client spend — company disclosures and news outlets beat vendor blogs).
  • Automation exposure by level and task, not by job title. The useful question is "which half of this job automates first".
  • Pay at 1–3, 5 and 10 years for each path, in Australia, with the role titles actually advertised.
  • Local market (the home city specialisation, employer set, whether the work is here or interstate).
  • The strongest counter-case to the answer you expect. Search for it deliberately. Run node tools/signals.mjs for the candidate's own market data: what the hunt has actually seen, which beats any national statistic for "what can I get locally this month".

2. Grade the sources, out loud

In the report, include a short source-quality warning: which findings come from recruiters, staffing firms or salary aggregators (commercial interest, soft magnitudes), which come from company disclosures, regulators, news outlets or academic work (harder), and where they disagree. Never present an aggregator's average as a fact the candidate can plan around. If the evidence is thin, say the direction is reliable and the magnitude is not.

3. Score the options

A table: rows = criteria (demand durability, automation exposure, optionality at 35, earnings at 5/10 years, local fit, uses what he already has, personal fit); columns = options. Prose in the cells, not scores out of ten — a fake number is worse than a sentence. Then the trajectory table: where each path puts him at 2, 5 and 10 years, in title, skills and pay.

4. Decide

State a recommendation in one sentence, then the case for it in one paragraph, then what it means concretely over the next 12 months as numbered actions tied to real things (live postings, a project, a certification, a conversation). Hedging is allowed only where it is honest: name the thing you genuinely cannot know.

5. Kill criteria and the human question

  • Kill criteria: the specific, observable events that would reverse the recommendation. Where possible make them measurable from this repo (see tools/signals.mjs) so the agent can watch them rather than the candidate having to remember.
  • The human question: end with what the analysis cannot answer and the cheapest experiment that would.

6. Record it

Append to profile/decisions.md: date, the decision, the recommendation, the two or three facts it rests on, the kill criteria, and a review date. On later runs, read that file first and say explicitly what has changed since the last decision — a decision log is only useful if it is revisited.

Hard rules

  • Never invent a statistic or a salary band. Cite or omit.
  • Never let a recruiter's number stand unqualified.
  • Say what you would do, and why, when asked. A refusal to answer is not neutrality; it is just less useful.
  • The goal is a decision the candidate can act on this week, not a survey of considerations.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.