benbrastmckie/nvim/skill-email-implementation

Email Implementation Skill

Thin wrapper that delegates email triage/cleanup implementation to the email-implementation-agent subagent. This skill never calls the wrapper binaries itself — it only prepares delegation context and dispatches.

Trigger Conditions

This skill activates when:

• Task type is "email"
• /implement command targets an email task
• A census/classify/archive/delete/unsubscribe-extract plan needs wrapper-only execution

Execution Flow

Stage 1: Input Validation

Validate task_number exists, task_type is "email", and an implementation plan is present.

Stage 2 + Stage 3: Preflight Status Update and Postflight Marker

Source skill-base.sh once, then follow @.claude/context/patterns/skill-preflight-flow.md in full for Stage 2 (preflight status update) and Stage 3 (marker creation):

source .claude/scripts/skill-base.sh
padded_num=$(printf "%03d" "$task_number")
skill_name="skill-email-implementation"
operation="implement"

Stage 3a: $PATH Precondition Check (contract §9)

Before dispatching, verify the five wrapper binaries are on $PATH:

command -v email-census email-classify email-archive-confirmed email-delete-confirmed \
  email-unsubscribe-extract

If any are missing, do NOT dispatch. Write status: "failed" metadata naming the missing binary and instructing the user to run home-manager switch --flake .#<user> to activate the generation containing modules/home/email/agent-tools.nix.

Stage 4a: Memory Retrieval and Literature Detection

Skip memory retrieval if: clean_flag is true (from --clean).

if [ "$clean_flag" != "true" ]; then
  memory_context=$(bash .claude/scripts/memory-retrieve.sh "$description" "$task_type" "" 2>/dev/null) || memory_context=""
fi

Follow @.claude/context/patterns/lit-stage4a-flow.md in full to resolve --lit and set lit_context, exactly as skill-orchestrate does. This skill supplies the shared block's preconditions: lit_flag, description, orchestrator_mode (default "false" when unset).

Stage 4: Prepare Delegation Context

Domain-specific context for email-implementation-agent:

• Wrapper-only contract: only the five named binaries, never raw himalaya/notmuch/msmtp/

secret-tool, never rm against Maildir.

• Propose-review-confirm-execute: STOP at every review gate for explicit human approval before

any --execute --confirm-manifest <sha256> call.

• Constants: MAX_BATCH_SIZE=50, PLAN_EXPIRY_DAYS=7, delete auto-propose confidence >= 0.90.

{
  "session_id": "sess_{timestamp}_{random}",
  "delegation_depth": 1,
  "delegation_path": ["orchestrator", "implement", "skill-email-implementation"],
  "timeout": 7200,
  "task_context": {
    "task_number": N,
    "task_name": "{project_name}",
    "description": "{description}",
    "task_type": "email"
  },
  "plan_path": "specs/{NNN}_{SLUG}/plans/MM_{short-slug}.md",
  "metadata_file_path": "specs/{NNN}_{SLUG}/.return-meta.json"
}

If memory_context and/or lit_context from Stage 4a are non-empty, include them in the prompt (memory context first, then literature briefing). Do NOT inject an empty block for either.

Stage 5: Invoke Subagent

Use Agent tool with subagent_type: "email-implementation-agent".

Stage 5b: Self-Execution Fallback

Follow @.claude/context/patterns/skill-self-execution-fallback.md in full. This skill's success status value for that block's write obligation is "implemented". This path is not expected for email tasks — the wrapper-only contract is agent-owned — but if it is reached, it must still respect Stage 3a's $PATH check and never invoke a wrapper binary directly from this skill.

Postflight (ALWAYS EXECUTE)

Stage 6: Parse Subagent Return

Read the metadata file from specs/{N}_{SLUG}/.return-meta.json, including memory_candidates.

Stage 7, 7a, 8, 8a, 9: Postflight Status, Memory Candidates, Artifact Linking, Notify, Cleanup

Follow @.claude/context/patterns/skill-postflight-flow.md in full: field_name=**Summary**, next_field=**Description**.

Stage 10: Return Brief Summary

Return 3-6 bullets. State clearly whether any mutation (--execute) occurred and, if so, which manifest/sha256 authorized it.

MUST NOT (Postflight Boundary)

After the agent returns, this skill MUST NOT:

1. Call any wrapper binary directly - All classify/archive/delete/unsubscribe execution is

done by the dispatched agent, wrapper-only, per the email extension's safety invariants

1. Run build/test commands - Verification is done by the dispatched agent
2. Use MCP/WebSearch tools - Domain tools are for the dispatched agent's use only
3. Analyze or grep source - Analysis is dispatched-agent work
4. Write manifests or reports - Artifact creation is dispatched-agent work

The postflight phase is LIMITED TO:

• Reading agent metadata file
• Calling update-task-status.sh for status updates (state.json + TODO.md)
• Linking artifacts in state.json
• Cleanup of temp/marker files

Reference: @.claude/context/standards/postflight-tool-restrictions.md