Agent skill

Find Me Skills

by luongnv89 in luongnv89/asm

Find Agent Skills for a goal the user cannot name yet, then export an installable bundle.

MITAuto-check passedAgent Workflows

Install Find Me Skills

skills CLI
$ npx skills add luongnv89/asm --skill find-me-skills -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install luongnv89/asm find-me-skills --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/luongnv89/asm.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/find-me-skills .claude/skills/find-me-skills && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
find-me-skills
GitHub stars
955
Token cost
~2.8k tokens
SKILL.md length
1,607 words
Files
5 (incl. references)
Skills in repo
7
Repo updated
First seen
Licence
MIT

At a glance

Find Agent Skills for a goal the user cannot name yet, then export an installable bundle.

  • Works in 6 steps: Collect intent → Confirm understanding (do not skip) → Discover candidates from the live catalog → …
  • They ask which skills fit a project
  • SKILL.md covers The loop, Prerequisite (check once, up…, Step 1 — Collect intent and Step 2 — Confirm understanding…, plus 8 more sections
  • Calls npm and bundle

What it does

Find Me Skills is an agent skill from luongnv89/asm. Find Agent Skills for a goal the user cannot name yet, then export an installable bundle. Use when they ask which skills fit a project. Don't use for installing named skills, authoring skills, or catalog maintenance.

Its SKILL.md is about 2.8k tokens, which your agent loads only when the skill is triggered. The skill folder holds 6 other files, including reference files (for example `evals/evals.json`, `references/bundle-format.md` and `references/catalog-discovery.md`). Compatibility notes: Claude Code with the asm CLI on PATH

It sits in Agent Workflows. The repository describes itself as: The universal skill manager for AI coding agents. The licence is MIT.

When your agent uses it

  • They ask which skills fit a project
  • Installing named skills
  • Authoring skills
  • Catalog maintenance

Example prompts

  • “/find-me-skills”

Requirements

  • Node.js
  • Compatibility (from SKILL.md): Claude Code with the `asm` CLI on PATH

Workflow steps

6 steps, taken from the step headings in SKILL.md.

  1. Collect intent
  2. Confirm understanding (do not skip)
  3. Discover candidates from the live catalog
  4. Curate: dedupe and explain
  5. Sequence into a step-by-step path
  6. Export an installable bundle (on approval)

What it can do on your machine

Read from SKILL.md and the folder at commit 19da710. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • npm
    • bundle

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use npm, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

  • Compatibility

    Claude Code with the `asm` CLI on PATH

    From compatibility in the SKILL.md frontmatter.

Context cost

Find Me Skills loads about 2.8k tokens when it runs, and up to ~3.6k if it reads all its reference files. Until then it costs about 58 tokens; SKILL.md has 1,607 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~58
When it runs · the whole SKILL.md, loaded when a task matches
~2.8k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~3.6k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from luongnv89/asm at commit 19da710, republished under its MIT licence (© luongnv89). 1,607 words, ~2,794 tokens.

Download SKILL.mdSave it as .claude/skills/find-me-skills/SKILL.md (or your agent's skills folder). This skill also uses 4 other files; get the full folder from GitHub.
name
find-me-skills
description
Find Agent Skills for a goal the user cannot name yet, then export an installable bundle. Use when they ask which skills fit a project. Don't use for installing named skills, authoring skills, or catalog maintenance.
compatibility
Claude Code with the `asm` CLI on PATH
license
MIT
effort
medium
metadata.version
1.4.0
metadata.author
Luong NGUYEN <luongnv89@gmail.com>

Find Me Skills

Help a user who has a goal but not a skill list find the right Agent Skills, explain what each one does, lay out an order to run them in, and — if they approve — hand them a single installable bundle file.

The user's defining trait is that they don't know what to ask for. Someone who already knows they want frontend-design should just run asm install …. This skill is for "I'm building an app and want to do marketing from scratch, but I don't know marketing or which skills exist." Draw out the goal, confirm it, then map it onto real, installable skills from the live catalog.

The loop

  1. Collect intent — conversationally draw out what the user is trying to achieve.
  2. Confirm understanding — play back your read of their situation; let them correct it before you search.
  3. Discover — query the live asm catalog for candidate skills (never guess skill names).
  4. Curate — dedupe, group by step, and explain each skill in one plain sentence.
  5. Sequence — give a step-by-step path with the input and output of each step.
  6. Export — on approval, write a bundle file and give the one-line install command.

Start at the step the user is already at. If they open with a rich goal ("I need SEO, a landing page, and launch copy for my SaaS"), confirm quickly and move to discovery. If they're vague ("help me market my app"), spend more time in steps 1–2. Never skip step 2: confirming before searching keeps recommendations relevant.

If the user asks for one capability ("is there a skill that converts PDFs?"), skip the loop: run one asm search "<term>" --available --json and give the matching asm install <installUrl> line.

Prerequisite (check once, up front)

This skill drives the asm CLI for discovery and produces a file it installs. Run this before promising recommendations:

bash
command -v asm || echo "MISSING"

If the output is MISSING, tell the user the skill needs the Agent Skill Manager CLI on PATH, point them at npm install -g agent-skill-manager, and stop.

Step 1 — Collect intent

Ask open questions, one or two at a time, until you can state the user's goal in a sentence. Useful prompts:

  • What are you building or working on right now?
  • What outcome do you want — a launched product, a written artifact, a faster workflow?
  • What part feels hardest or most unfamiliar? (This is often where skills help most.)
  • Is this a one-off task or something you'll repeat?

Match your vocabulary to theirs. A non-marketer asking for "marketing" may actually need positioning, a landing page, and launch copy — offer those as options; do not assume. Avoid jargon ("ICP", "ASO") unless they use it first.

Step 2 — Confirm understanding (do not skip)

Before searching, play back what you heard and ask for an explicit confirmation:

Here's what I understand: you're building {project}, and you want to {goal}. The pieces you're unsure about are {gaps}. Did I get that right?

If they correct you, fold it in and ask again. Move on only after they agree. Confirming the situation is what prevents a confidently-wrong skill list.

Step 3 — Discover candidates from the live catalog

Never invent skill names or install URLs. The catalog changes constantly; the only trustworthy source at runtime is the asm CLI on the user's machine. Derive 2–5 search terms from the confirmed goal and query each:

bash
asm search "<term>" --available --json

Run a separate search per term — broad terms ("marketing", "seo", "landing page", "launch") surface different skills. Read references/catalog-discovery.md when you need the JSON shape, installed-vs-available rules, or empty-result handling; keep this detail out of the main context budget until Step 3 needs it.

Read each candidate's description to judge relevance; its "Use when…" / "Don't use for…" text tells you whether it fits the goal. Then run asm search "<term>" --json without --available to detect installed skills. Mention installed matches in the plan, but exclude them from the bundle. Only available skills with an installCommand go in.

Step 4 — Curate: dedupe and explain

From the union of search hits, build the recommendation set:

  • Deduplicate by skill name. Keep one entry per name. If two repos offer the same name, keep the one whose description best matches the goal and note the choice in the plan.
  • Drop weak fits. Keep only skills you can justify in one sentence tied to the user's goal.
  • Explain each in plain language — one sentence on what it does for this user, not a paraphrase of its description. "landing-page-copywriter writes the words for your launch page so visitors understand and sign up."

Step 5 — Sequence into a step-by-step path

Order the curated skills into the sequence the user should run them in. For each step, state its input (what the user or the previous step provides) and its output (what they'll have after). Foundational/context skills usually come first; review/QA skills usually come last. Example shape:

Step 1 — marketing-context
  in:  your product description, target customer
  out: a saved brand/positioning brief other skills read first
Step 2 — landing-page-copywriter
  in:  the brief from step 1
  out: landing-page copy ready to paste

Show this plan to the user before exporting anything. Make it easy to say "drop step 2" or "add something for email". After any change, show the revised plan and ask for approval again.

Step 6 — Export an installable bundle (on approval)

Only after the user approves the plan, write a bundle file in asm's BundleManifest format with a goal-based name such as marketing-starter.bundle.json. Write it to the current directory unless the user names another path.

See references/bundle-format.md for the required JSON template, validation rules, and the reason this skill uses asm bundle install instead of asm install or asm import. Copy each installUrl from Step 3 verbatim; never hand-construct it and never include already-installed skills.

Check the file before handing it off:

bash
asm bundle show ./marketing-starter.bundle.json --json
  • If the command exits 0 and its skills[].name list matches the approved plan, give the user the install command asm bundle install ./marketing-starter.bundle.json. In a terminal it asks for the tool, then which bundle skills to install, then the scope. Outside a terminal, -p/--tool <tool> is required; -y skips the skill and scope pickers.
  • If the command errors, fix the field the error names, then run the check again. After two failed checks, stop, report blocked with the error, and do not hand off an install command.
Show full SKILL.md (589 more words)Show less

Step Completion Reports

After each step, print a compact status block: √ pass, × fail, — context.

◆ Step N — [step name]
··································································
  [check]:           √ pass
  Result:            PASS | FAIL | PARTIAL

Checks per step:

  • Prerequisite — asm on PATH
  • Steps 1–2 — goal stated in one sentence, user confirmed the goal
  • Step 3 — one available search per term, one installed search per term
  • Steps 4–5 — no duplicate names, every step has in/out, user approved the plan
  • Step 6 — bundle file written, asm bundle show --json exit 0, listed skills match the plan

Final output

End every run with a short report the user can read without scrolling back:

  1. Result — the first line states complete, plan only (nothing to bundle, or the user declined export), or blocked (with the reason).
  2. The plan — numbered steps, each with the skill, a one-line purpose, and in/out.
  3. Evidence — the search terms you ran, the bundle file path, and whether asm bundle show passed.
  4. Uncertainty — say that each recommendation is based on the skill's catalog description; you did not run, test, or security-audit any skill. Name any part of the goal with no catalog match.
  5. Decision — for complete, the user's next action: asm bundle install ./<file> on its own line; installation happens only when they run it, so no other approval is needed. For plan only or blocked, name what would unblock the next step, or state that no action is needed.

Keep explanations plain. The user came here because they didn't know the landscape — leave them understanding what they're about to install and why.

Acceptance Criteria

Verify all of these before calling the run complete:

  • The user explicitly confirmed the goal before any catalog search.
  • At least one asm search "<term>" --available --json query was run for each chosen search term.
  • Installed skills were checked with asm search "<term>" --json and excluded from the bundle.
  • The user approved the sequenced plan before any file was written.
  • Expected output for complete: a numbered plan, a bundle file path, and an install command. For plan only: the numbered plan and no bundle file.
  • For complete, asm bundle show ./<file>.bundle.json --json succeeded; otherwise the run reports blocked with the validation error instead of handing off a broken command.

When reviewing a run's output (by eval or by a human), also check that a reader can:

  • Find the main result — the first line states the completion status.
  • Separate facts from assumptions — catalog-description judgements and untested behavior are labeled as such.
  • Trace claims — each recommended skill maps to a search term and a catalog result; "bundle valid" maps to the asm bundle show check.
  • See the next decision — the install command is named, or the report states why there is none.

An agent's own review cannot confirm human understanding. A reviewer who received no human feedback records understanding as unconfirmed.

Edge cases

  • asm not installed — stop at the prerequisite check; point them at the install docs.
  • Vague goal that won't sharpen — stay in steps 1–2; offer 2–3 concrete directions ("Do you mean A, B, or C?") rather than searching on a guess.
  • No catalog matches for part of the goal — say so; recommend only what fits, and suggest skill-creator if they may need to author something that doesn't exist yet.
  • Everything relevant is already installed — there's nothing to bundle; give the step-by-step plan using their installed skills, skip the export, and report the result as plan only.
  • User declines the plan — don't write a file. Adjust from their feedback and ask again, or stop and report plan only.
  • Duplicate skill names across repos — keep one; pick the better-matching description and note the choice.

© luongnv89, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 4 other files (references) in skills/find-me-skills of luongnv89/asm.

  • SKILL.md
  • LICENSE.txt
  • evals/evals.json
  • references/bundle-format.md
  • references/catalog-discovery.md

Open the folder on GitHubat commit 19da710

Compare with similar skills

Find Me Skills next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Find Me Skills compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Find Me Skills this skillluongnv89/asm955—~2.8kAutomated safety check: PassMIT
MCP Server Builderanthropics/skills180k63 repos~2.3kAutomated safety check: PassApache-2.0
Hook Development for Claude Code Pluginsanthropics/claude-plugins-official38k10 repos~4.1kAutomated safety check: NotesApache-2.0
Using Superpowersfarm-fe/farm5.6k36 repos~1.4kAutomated safety check: PassMIT
Executing Plans Inlineobra/superpowers297k2 repos~5.1kAutomated safety check: PassMIT
Skill CreatorAzure/azqr79689 repos~8.2kAutomated safety check: PassApache-2.0

Similar skills

  • MCP Server Builder

    anthropics/skills

    Official

    Guides the design and implementation of Model Context Protocol servers in TypeScript or Python, from tool naming and error messages to evaluation.

    180k GitHub starsUsed in 63 repos~2.3k tokens
    Agent WorkflowsAuto-check passed
  • Hook Development for Claude Code Plugins

    anthropics/claude-plugins-official

    Official

    Explains how to write Claude Code plugin hooks, both prompt-based checks and bash commands, for events such as PreToolUse, Stop and SessionStart.

    38k GitHub starsUsed in 10 repos~4.1k tokens
    Agent WorkflowsAuto-check: notes
  • Using Superpowers

    farm-fe/farm

    A skill your agent uses when starting any conversation - establishes how to find and use skills, requiring Skill tool invocation before ANY response including clarifying questions

    5.6k GitHub starsUsed in 36 repos~1.4k tokens
    Agent WorkflowsAuto-check passed
  • Executing Plans Inline

    obra/superpowers

    Has the agent carry out an implementation plan itself, task by task in the current session, keeping a ledger, proving each step with a test and ending with one whole-branch review.

    297k GitHub starsUsed in 2 repos~5.1k tokens
    Agent WorkflowsAuto-check passed
  • Skill Creator

    Azure/azqr

    Official

    Create new skills, modify and improve existing skills, and measure skill performance.

    796 GitHub starsUsed in 89 repos~8.2k tokens
    Agent WorkflowsAuto-check passed
  • Claude Code Agent Development

    anthropics/claude-plugins-official

    Official

    Explains how to write agents for Claude Code plugins: the markdown file with YAML frontmatter, trigger descriptions, model and color settings, and system prompt design.

    38k GitHub starsUsed in 7 repos~2.8k tokens
    Agent WorkflowsAuto-check passed

More from luongnv89/asm

  • Skill Shortener

    luongnv89/asm

    Refactor a too-long SKILL.md by progressive disclosure: measure token cost, classify every section KEEP/CUT/MOVE, shorten the body into references/ and scripts/, verify nothing was lost.

    955 GitHub stars~3.8k tokensUpdated today
    Auto-check: notes
  • Skill Creator

    luongnv89/asm

    Create a skill or bring an existing one up to the same standard (validate + asm eval fix loop); run evals, tune triggering.

    955 GitHub stars~5.3k tokensUpdated today
    Auto-check passed
  • Refresh Index

    luongnv89/asm

    Sync every enabled repo in the curated skill index and open a confirmation-gated PR.

    955 GitHub stars~3.4k tokensUpdated today
    Auto-check: notes
  • Skill Index Updater

    luongnv89/asm

    Add GitHub skill repos to the ASM index: clone, audit, eval, regenerate index, rebuild catalog, open PR.

    955 GitHub stars~4.3k tokensUpdated today
    Auto-check: notes
  • Install an improved variant of one named skill: resolve it by local path, repo, or name, run skill-creator's retrofit on a throwaway copy, then install the improved result.

    955 GitHub stars~3.3k tokensUpdated today
    Auto-check: notes
  • Skill Upstream PR

    luongnv89/asm

    Improve an open-source GitHub skill and open a friendly suggestion PR upstream: fork, run skill-creator's retrofit, attach asm eval before/after metrics.

    955 GitHub stars~2.4k tokensUpdated today
    Auto-check: notes

Categories

Questions about Find Me Skills

What does Find Me Skills do?

Find Agent Skills for a goal the user cannot name yet, then export an installable bundle. Find Me Skills is an agent skill from luongnv89/asm. Find Agent Skills for a goal the user cannot name yet, then export an installable bundle.

When should I use Find Me Skills?

Find Me Skills fits situations like: they ask which skills fit a project; installing named skills; authoring skills; catalog maintenance.

How do I install Find Me Skills in Claude Code?

Run `npx skills add luongnv89/asm --skill find-me-skills -a claude-code`. Or copy the skill folder (skills/find-me-skills in luongnv89/asm) into .claude/skills/find-me-skills in your project. Claude Code loads it when a task matches its description.

How do I install Find Me Skills in Codex?

Run `npx skills add luongnv89/asm --skill find-me-skills -a codex`. Or copy the skill folder (skills/find-me-skills in luongnv89/asm) into .agents/skills/find-me-skills in your project. Codex loads it when a task matches its description.

Can I use Find Me Skills in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add luongnv89/asm --skill find-me-skills -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/find-me-skills, .gemini/skills/find-me-skills, .github/skills/find-me-skills and .opencode/skills/find-me-skills in your project.

What does Find Me Skills need to run?

Going by SKILL.md and its folder, Find Me Skills needs the command-line tools its instructions call (npm and bundle). Our summary lists: Node.js. Compatibility (from SKILL.md): Claude Code with the `asm` CLI on PATH.

Does Find Me Skills access the network?

SKILL.md contains no URLs. Its commands use npm, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Find Me Skills safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Find Me Skills use?

Find Me Skills is published under the MIT licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Find Me Skills use?

About 2.8k tokens (SKILL.md is roughly 11k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 775 tokens, read only when the agent opens those files.

What are the alternatives to Find Me Skills?

Skills that share tags, products or a category with Find Me Skills: MCP Server Builder (anthropics/skills, 180k stars), Hook Development for Claude Code Plugins (anthropics/claude-plugins-official, 38k stars), Using Superpowers (farm-fe/farm, 5.6k stars) and Executing Plans Inline (obra/superpowers, 297k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Find Me Skills?

luongnv89 (a GitHub user) maintains it in luongnv89/asm, which has 955 GitHub stars. The repository holds 7 skills in this directory. The repository was last updated on October 10, 2026.

Source: luongnv89/asm on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.