suggest-roles

Invoke both honest-advisor and market-researcher to generate role suggestions and write actionable signals that improve job-scout ranking.

Important invocation rule:

• Use the exact agent names honest-advisor and market-researcher.
• Do not invent or alias agent types (for example, do not call "career-assessment" or "market-analysis" agent types).
• If either agent invocation fails, retry once using the exact names above before returning an error.

Workflow

1. Confirm data exists

Read:

• {user_dir}/CareerNavigator/profile.md
• {user_dir}/CareerNavigator/ExperienceLibrary.json

If profile has no target roles:

"I need your current target role(s) first. Run /career-navigator:launch or update CareerNavigator/profile.md before running role suggestions."

If ExperienceLibrary units is empty:

"Your ExperienceLibrary is empty. Run /career-navigator:add-source to add a resume before role suggestions."

Optionally read {user_dir}/CareerNavigator/tracker.json for confidence and outcome context.

2. Run advisor pass (competitiveness + transferable fit)

Invoke honest-advisor in assessment mode for the user's primary target role. Ask it to:

• identify under-covered requirements
• identify nearby role variants where the user's strongest signals are more competitive
• return 3-6 candidate roles with rationale

3. Run market pass (demand + displacement + geography)

Invoke market-researcher for the same role set and geography. Ask it to:

• classify demand posture for each candidate role (rising/stable/softening)
• flag displacement risk posture
• identify geography-specific competitiveness constraints/opportunities

4. Synthesize suggestions

Combine both outputs into a ranked role list:

• prioritize roles where transferable fit is strong and demand posture is favorable
• down-rank roles with weak fit, softening demand, or severe geography mismatch
• include at least one "stretch but plausible" option if evidence supports it

5. Write job-scout scoring signals

Update {user_dir}/CareerNavigator/tracker.json with a strategy_signals object (create if missing):

{
  "strategy_signals": {
    "as_of": "YYYY-MM-DD",
    "recommended_role_types": ["..."],
    "adjacent_role_types": ["..."],
    "deprioritize_role_types": ["..."],
    "market_tailwinds": ["..."],
    "market_headwinds": ["..."],
    "preferred_geographies": ["..."],
    "avoid_geographies": ["..."],
    "signal_confidence": "Preliminary | Directional | Moderate | High",
    "source": ["honest-advisor", "market-researcher"]
  }
}

These signals are consumed by job-scout to improve ranking.

6. Present result

Format:

**Suggested Roles** — {today's date}
Confidence: {tier}

Top role opportunities
1. {Role}
   - Why this fits: {transferable-fit rationale}
   - Market signal: {demand/geography/displacement summary}
2. ...

Deprioritize for now
- {Role}: {why}

Job-scout update
- Recommended role types written: {list}
- Deprioritized role types written: {list}
- Geography signals written: {list}

7. Suggest next step

After presenting:

• Suggest /career-navigator:search-jobs to run ranking with the updated role/geography strategy signals.
• Suggest /career-navigator:tailor-resume for the top suggested role.

Failure fallback:

• If one agent still fails after retry, continue with the successful agent output, clearly label partial-completion status, and tell the user which exact agent failed.