Doomersion language feed

Workflow

When this skill triggers, follow these steps in order.

Step 1 — Capture Input Parameters

Ask the user for (or read from the prompt):

1. Target language — the language to learn (e.g. Spanish, French, Japanese, Portuguese).
2. Proficiency level — CEFR level: A1, A2, B1, B2, C1, or C2.
3. Interests — one or more topics the learner cares about (e.g. football, cooking, travel, music, gaming).
4. Feed length — how many posts to generate (default: 10; minimum: 8).

If any required field is missing, prompt:

"Please tell me: (1) the language you want to learn, (2) your CEFR level (A1–C2), (3) your interests, and (4) how many posts you'd like (default 10)."

Step 2 — Plan Vocabulary and Grammar Coverage

Before writing posts, internally plan:

• Select N vocabulary items appropriate to the stated CEFR level (N ≈ feed length × 1.5 so no two posts repeat the same word).
• Identify 2–3 grammar structures suitable for the level:
  • A1/A2: present tense, articles, basic adjective agreement, ser/estar (Spanish), negation.
  • B1/B2: subjunctive mood, compound tenses, relative clauses, passive voice.
  • C1/C2: complex subordination, idiomatic expressions, register variation.
• Map each post to one vocabulary item and one grammar structure so the feed has variety.

Step 3 — Generate Each Post

For each of the N posts, produce the following block in order:

**@[handle]** [emoji]
[Target-language post text — 1–4 sentences on the interest topic]

> [English translation / gloss]

**Note:** [One grammar or vocabulary explanation — concise, ≤2 sentences]

---

Rules for each post:

• The post body must be entirely in the target language — no English in the post body.
• Vocabulary and sentence complexity must match the stated CEFR level.
• The topic must relate to the stated interest(s).
• The handle should be a plausible social-media username related to the interest (e.g. @golazo_fc, @futbol_diario).
• The emoji should be thematically appropriate.
• The English gloss should be a natural translation, not word-for-word.
• The grammar/vocab note should highlight the most pedagogically useful feature of that post (new word, conjugation pattern, false cognate, idiomatic phrase, etc.).

Step 4 — Format the Complete Feed

Output the posts as a scrollable feed under a header:

# [Language] Feed — [Level] · [Interest(s)]

Follow all posts with a ## Rubric section that lists the quality criteria this artifact satisfies (see example).

Step 5 — Quality Check

Before finalising, verify:

• At least the requested number of posts are present (minimum 8).
• Every post has: target-language body, English gloss, one grammar/vocab note.
• No English appears in the post body.
• Vocabulary complexity matches the stated CEFR level.
• All posts relate to the stated interest(s).
• No vocabulary item is repeated across posts (same root word counts as a repeat).
• The ## Rubric section is present.

If any check fails, revise before returning the output.

Example

See examples/input.md for a worked request (Spanish A2, football interest, 10 posts) and examples/output.md for the corresponding generated feed.