agentleFS
Sign inStart free
liknenSliknenS/language-learning-skills/skills/speaking-partner/SKILL.md

speaking-partner

Skill0 starsChanged 25 days agoMITView on GitHub

Speaking partner for a foreign language - runs a role-play with a task (job interview, landlord, doctor, small talk, a meeting) at the learner's level, with target phrases and a short review of the learner's lines afterwards. Use this skill whenever a language learner wants to talk, practise a conversation, role-play a situation or rehearse an interview in the language they study. Also for "давай поговорим по-английски", "поговорімо польською", "porozmawiajmy".

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. nexaforgelab/english-skill★ 4same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository5 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 5 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    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:

    Push a little: If the learner answers in one word, ask a follow-up that needs a full sentence.

    line 17
  2. Distinctive:

    Review: After 6-12 turns stop and give: what worked, up to three corrections (their line → a natural version), one or two upgrades a native …

    line 19
  3. Very specific:

    Start with today() and profile() if not done this session and read docs("talk").

    line 28
  4. Distinctive:

    No window (some hosts, phones): the tool reply has an app link to continue the conversation there.

    line 39
  5. Distinctive:

    Run the conversation in text, or in your host's voice mode if it has one.

    line 43
  6. Distinctive:

    Keep your notes of mistakes during the talk and give the review at the end.

    line 45

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “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

Speaking partner

A conversation with a task, not small talk for its own sake.

Steps

  1. Set the scene. Role, place, what the learner needs to achieve ("rent a flat, ask about the deposit and when you can move in"). Agree on 3-5 target phrases the learner should use.
  2. Play the role. Short turns, at the learner's level, one question at a time. Stay in the language being learned. Do not correct during the talk unless the learner asks for live hints; note mistakes for later.
  3. Push a little. If the learner answers in one word, ask a follow-up that needs a full sentence. Bring the target phrases up naturally so the learner has a reason to say them.
  4. Review. After 6-12 turns stop and give: what worked, up to three corrections (their line → a natural version), one or two upgrades a native speaker would say, and which target phrases were used.
  5. Keep what is worth it. Offer to save the corrections and upgrades as vocabulary (see vocab-coach).

Scene ideas by goal: work (interview, stand-up, client call), relocation (bank, landlord, residence office), daily life (pharmacy, delivery at the door, neighbour), exam (examiner questions in the exam format, see exam-prep).

If Talkfirst MCP is connected

  • Start with today() and profile() if not done this session and read docs("talk"). Weave today().targets into the conversation.
  • dialogue_start opens a voice conversation window: the learner speaks, an AI partner answers aloud, and each line gets a review (corrections, rephrasings, pauses before words). Pass brief (role, scene, what to ask, up to 2000 characters), targets (3-5 phrases: strings already in the vocabulary, new ones as {term, meaning}) and expectations (what counts as success). hints="live" shows reviews right away, end after the talk.
  • On the free plan an own topic and brief are not available, only ready scenes: call dialogue_start(scene=...); if you do not know the scene ids, the error lists them. There is also a limit on conversations started from the chat per hour.
  • The window sends the result itself; late reviews come via dialogue_result(session_id), past talks via dialogue_history(). Use the result for the review step instead of your memory of the talk.
  • No window (some hosts, phones): the tool reply has an app link to continue the conversation there.

Without it

Run the conversation in text, or in your host's voice mode if it has one. Voice mode transcribes generously, so hesitations and guessed words get lost: ask the learner to say which words they had to search for, and review those. Keep your notes of mistakes during the talk and give the review at the end.

Voice conversations with per-line reviews and pause tracking, kept on the learner's account: connect Talkfirst (free plan, MCP connector for Claude and ChatGPT).

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.

Check your setup, free