Resume Ops

Read the config file. Remember the resolved path for later use (Step 7). Validate that required fields are present:

• user.name
• user.github_username
• At least one source with enabled: true
• Warn (but don't fail) if user.target_role or role_history is missing — these improve synthesis quality

Step 2: Check Prerequisites

If sources.github.enabled is true:

gh auth status

If this fails, tell the user to run gh auth login first.

If sources.jira.enabled is true:

• Check if Atlassian MCP tools are available
• If not, warn the user and skip Jira gathering

Step 3: Create Output Directory

mkdir -p resume-ops-output/raw

Step 4: Spawn Parallel Subagents

Launch subagents in parallel using the Agent tool. For each enabled source, spawn one agent.

GitHub Agent (if sources.github.enabled):

Read the file agents/github.md from the resume-ops skill directory. Replace all {{template_vars}} with values from resume-ops.yaml:

• {{github_username}} → user.github_username
• {{orgs}} → sources.github.orgs (as comma-separated list)
• {{since}} → sources.github.since
• {{include_reviews}} → sources.github.include_reviews
• {{exclude_repos}} → sources.github.exclude_repos (as comma-separated list)

Dispatch with the Agent tool:

Agent(
  description: "Gather GitHub contributions",
  subagent_type: "general-purpose",
  prompt: [filled github.md template]
)

Jira Agent (if sources.jira.enabled):

Read the file agents/jira.md from the resume-ops skill directory. Replace all {{template_vars}}:

• {{assignee}} → sources.jira.assignee
• {{project_keys}} → sources.jira.project_keys (format as comma-separated quoted strings for JQL, e.g., "PLAT", "CDC")
• {{since}} → sources.github.since (reuse the same date boundary)
• {{ignore_labels}} → sources.jira.ignore_labels

Dispatch with the Agent tool:

Agent(
  description: "Gather Jira contributions",
  subagent_type: "general-purpose",
  prompt: [filled jira.md template]
)

IMPORTANT: Launch both agents in a SINGLE message with multiple Agent tool calls so they run in parallel.

Step 5: Verify Raw Data

After both agents complete, verify output files exist:

ls -la resume-ops-output/raw/

Read each raw file and check it's not empty or error-only. If a source failed:

• Log the failure
• Continue with available data

Step 6: Run Synthesis

Read templates/synthesis.md from the resume-ops skill directory. Replace template vars:

• {{name}} → user.name
• {{target_role}} → user.target_role
• {{role_history}} → formatted role_history from config
• {{focus_areas}} → synthesis.focus_areas
• {{highlight_keywords}} → synthesis.highlight_keywords
• {{max_bullets_per_theme}} → synthesis.max_bullets_per_theme
• {{team_size}} → synthesis.team_size

Then read ALL raw data files and follow the synthesis instructions to generate resume-ops-output/resume-bullets.md.

This step runs inline (not as a subagent) so you have full context of the raw data.

Step 7: Copy Config Snapshot

Copy the config file that was found in Step 1 (may be in cwd or ~/.config/resume-ops/):

cp [resolved_config_path] resume-ops-output/resume-ops.yaml

Use the actual path where you found the config, not a hardcoded relative path.

Step 8: Present Results

Tell the user:

resume-ops complete! Here's what was gathered:

• GitHub: [X merged PRs, Y reviews, Z repos]
• Jira: [X tickets, Y epics, Z critical issues]

Output files:

• resume-ops-output/raw/github.md — raw GitHub data
• resume-ops-output/raw/jira.md — raw Jira data
• resume-ops-output/resume-bullets.md — synthesized resume bullets

Next step: Review resume-bullets.md and fill in any TODO markers with specific metrics from your memory or dashboards.

Error Handling

• If no config found → clear message with setup instructions
• If gh CLI not authenticated → tell user to run gh auth login
• If Atlassian MCP not available → skip Jira, warn user, continue with GitHub only
• If a subagent fails → log error, continue with other sources
• If no raw data at all → tell user to check config and prerequisites