Everything you need to start, progress and complete.كل ما تحتاج إليه للبدء والتقدّم وإكمال الدورات

Outcome:
AI may:
AI must not:
Human decides:
Evidence required:
Stop or escalate when:

Claim:
Claim type: fact | interpretation | recommendation
Primary source:
Source date:
Check performed:
Verdict: supported | uncertain | rejected

Task and audience:
Verified context:
Constraints:
Required output:
Evaluation criteria:
What changed from v1 to v2:
Measured result:

[ ] Facts verified
[ ] Personal data removed or authorised
[ ] AI contribution reviewed or disclosed
[ ] Risks and limitations recorded
[ ] Human owner approved the final result

Read AGENTS.md, the package scripts and the relevant source tree.
Do not edit anything.
Explain:
1. where this feature lives,
2. how it is tested,
3. which rules constrain a change,
4. what remains uncertain.

Plan only: add an Arabic/English toggle to this page.
Name the files and accessibility behaviour.
Wait for approval.
Then change only approved files, run focused tests,
and summarise the diff and remaining risks.

Diagnose this failing test without editing yet.
Show:
- the observed failure,
- two plausible causes,
- evidence for the most likely cause,
- the smallest safe fix,
- tests that must pass afterward.

Refactor only the duplicated formatter in the two named files.
Preserve exports and output exactly.
Do not rename unrelated code.
Run the existing tests, add one edge-case test,
then show the complete diff for review.

const result = await client.messages.create({
  model: "claude-opus-5",
  max_tokens: 600,
  system: "Return Arabic action items. Mark unknown facts clearly.",
  messages: [{ role: "user", content: transcript }],
});

const facts = JSON.stringify(eventRecord);
const result = await client.messages.create({
  model: "claude-opus-5",
  max_tokens: 700,
  system: "Write Arabic and English copy; never alter supplied facts.",
  messages: [{ role: "user", content: facts }],
});

const result = await client.messages.create({
  model: "claude-opus-5",
  max_tokens: 400,
  messages: [{ role: "user", content: [
    { type: "image", source: { type: "base64", media_type: "image/jpeg", data: receiptBase64 } },
    { type: "text", text: "Extract only visible receipt facts." },
  ]}],
});

const result = await client.messages.create({
  model: "claude-opus-5",
  max_tokens: 500,
  messages: [{ role: "user", content: JSON.stringify(application) }],
  output_config: { format: { type: "json_schema", schema: eligibilitySchema } },
});

const matches = await searchCatalog(query);
const result = await client.messages.create({
  model: "claude-opus-5",
  max_tokens: 500,
  system: "Use only the supplied catalog records. Say when no match exists.",
  messages: [{ role: "user", content: JSON.stringify({ query, matches }) }],
});

const result = await client.messages.create({
  model: "claude-opus-5",
  max_tokens: 400,
  cache_control: { type: "ephemeral" },
  system: glossary,
  messages: [{ role: "user", content: question }],
});

const result = await client.messages.create({
  model: "claude-opus-5",
  max_tokens: 700,
  tools: [{ type: "web_search_20260209", name: "web_search" }],
  messages: [{ role: "user", content: digestRequest }],
});

const result = await client.messages.create({
  model: "claude-opus-5",
  max_tokens: 500,
  tools: [{ name: "track_shipment", description: "Read shipment status",
    strict: true, input_schema: trackingSchema }],
  messages: [{ role: "user", content: question }],
});