goaly-mission

name	goaly-mission
description	Use when doing weekly planning (Mondays) or daily pulse checks. Two modes: MONDAY (full planning with retro + energy budget + task ranking), PULSE (weekday scorecard + flags). Triggers: '/mission', 'plan my week', 'weekly planning', 'quick check'.
allowed-tools	["Read","Edit","Write","Grep","Glob","Bash (includes qmd CLI for semantic search)","AskUserQuestion"]

/goaly-mission — Strategic Mission

Before You Start

1. Read .claude/skills/_shared/conventions.md — shared process rules and data integrity gotchas
2. Read gotchas.md in this directory — mission-specific failure modes

Learned Preferences

!tail -20 "${CLAUDE_SKILL_DIR}/data/feedback.log" 2>/dev/null || echo "No feedback yet — preferences will accumulate over sessions."

Context

• Today: !date +%Y-%m-%d' ('%A')'
• Uncommitted changes: !git status --short notion-mirror/

Recent Runs

!tail -5 "${CLAUDE_SKILL_DIR}/data/run-log.txt" 2>/dev/null || echo "No previous runs."

Top-down strategic dashboard. KPIs first, patterns second, tasks last. Detects mode from day of week. Three-phase: collect data in parallel, analyze without I/O, interact with [Owner].

Coaching prep: Use /goaly-coaching-prep for [Coach] sessions (not mission).

STEP 0: Freshness Check

See _shared/conventions.md for sync verification and freshness check process.

STEP 1: Context Detection

day_of_week=$(date +%u) — 1=Monday, 7=Sunday

Mode	Condition	Steps
PULSE	Weekday (default)	0-5, Today
MONDAY	day_of_week == 1, or explicit planning trigger	All (0-10)

Announce: ## Mission — [Mode] ([Day, Date])

STEP 2: Parallel Data Collection

All I/O happens here. Fire all three groups simultaneously. No analysis — just gather.

Group A — qmd Semantic Discovery

• A1: qmd query "KPI progress targets current value tracking" (Step 4)
• A2: qmd query "stalled tasks blocked waiting initiation avoidance" (Step 5)
• A3: qmd query "client follow-up dormant inactive no recent contact" (Step 5)
• A4 (MONDAY): qmd query "completed shipped done accomplished this week" (Step 8)

Group B — Grep + Read

Fire in parallel, then Read each matching file:

• B1: Grep "^lifecycle: Active" notion-mirror/kpis/ (Step 4)
• B2: Grep "^status: Active" notion-mirror/clients/ (Steps 5/6)
• B3: Grep "^status: Done This Week" notion-mirror/tasks/ (Steps 3/8)
• B4: Grep "^status: Done This Week" notion-mirror/personal-tasks/ (Step 3)
• B5: Grep "^status: Planned this week" notion-mirror/tasks/ (Steps 5/8/9)
• B6: Grep "^status: In progress" notion-mirror/tasks/ (Steps 5/9)
• B7 (MONDAY): Grep "^status: Not started" notion-mirror/tasks/ (Step 9)
• B8: Grep "^energy: Deep Work" notion-mirror/tasks/ (Step 7)
• B9: Per active client from B2: Grep interactions, extract latest date. ALSO run gog gmail search 'from:CLIENT_EMAIL OR to:CLIENT_EMAIL newer_than:30d' per client to catch unlogged email contact. Use the most recent date from either source. (Step 5)
• B10: Grep "## Coaching Accountability" notion-mirror/tasks/ — finds all frog tasks with coaching tracking sections (Step 5/9)

Group C — External Systems

• C1: gog calendar events --all --account [e-mail] --plain (Steps 7/8)
• C2: gog gmail search 'newer_than:7d' --max 30 --account [e-mail] --plain (MONDAY) or 'newer_than:2d' --max 20 (PULSE) — apply triage rules from MEMORY.md
• C3 (MONDAY): Git log scans across Goals & Tasks, dev/client/, dev/frontend/ (7 days)

Also Read MEMORY.md for MRR computation and Active Leads.

STEP 3: Auto-Cleanup (MONDAY only)

Using B3 + B4: edit each "Done This Week" task to status: "Done". Commit: chore: archive Done This Week tasks. Report count.

STEP 4: Strategic Scorecard (all modes)

Uses: A1, B1, B2, C1, MEMORY.md. No new I/O.

KPI	Source	Method
MRR	MEMORY.md Active Leads	Sum monthly estimates (NOT rate x hours). Ask [Owner] to confirm.
Revenue Sources	B2 clients	Count clients with rate > 0
Weekly Conversations	C1 calendar	Count distinct external meetings past 7 days. Exclude baseline calendar.

Traffic Light Table: Light | KPI | Current | Target | Progress | Gap | Last Updated. Thresholds: Green >= 75%, Yellow 25-74%, Red < 25%. Staleness: Weekly >10d, Monthly >35d, Quarterly >100d.

MONDAY: AskUserQuestion for non-auto KPIs. Always ask for MRR confirmation. PULSE: No prompts unless critically stale (2x overdue).

STEP 5: Pattern Detection (all modes)

Uses: A2, A3, B1, B2, B5, B6, B7, B9, MEMORY.md. No new I/O.

Check	Flag	Detection
Portfolio Concentration	Any client > 50% MRR, or < 3 revenue sources	B2 + MEMORY.md
KPI Staleness	_notion_edited overdue vs tracking_frequency	B1
Initiation Avoidance	_notion_edited > 14 days on active tasks	A2 + B5/B6
Killed Mammoth	No interaction 14+ days, or accepted lead no follow-up	A3 + B9 + MEMORY.md
Frog Tasks	Planned/Not started, _notion_edited > 14 days	B5/B7, weeks = (today - edited) / 7
Frog Streak	Coaching-tagged tasks still "Not started" with sessions-committed > 1	B10 — count tasks with "Coaching Accountability" section + status "Not started". Show "X frog tasks stalled across Y coaching sessions"

Present all flags together. If none: "No flags. Portfolio, KPIs, and pipeline all healthy."

Today (PULSE only): Show In progress tasks (B6) as table: Task | Project | Energy. Then actionable emails from C2.

STEP 5b: Daily Frog (all modes)

Uses: B10. No new I/O.

Surface ONE frog task for today. Selection priority:

1. Highest sessions-committed count (most overdue coaching commitment)
2. If tied, oldest _notion_edited date (longest untouched)
3. If energy mismatch (Deep Work frog on a meeting-heavy day), pick next eligible

Present as:

🐸 TODAY'S FROG: [task title]
   Committed: [N] coaching sessions ([dates])
   Days stalled: [N]
   First step: [from "First micro-step" in task body]

MONDAY only: After task ranking (Step 9), prompt [Owner] to add calendar blocks for top 3 frogs this week using AskUserQuestion.

PULSE only: Check if yesterday's frog was eaten (status changed from "Not started" to anything else since last pulse). Track in run log as frog_eaten=true|false.

STEP 6: Portfolio Concentration (all modes)

Uses: B2, MEMORY.md. Reuse Step 5 data — no re-reading.

Always show: Client | Status | Type | Est. Monthly | Share. Revenue from MEMORY.md Active Leads, not rate fields.

STEP 7: Energy Budget (MONDAY only)

Uses: C1, B8, B5, B6.

1. Remaining weekdays (including today if Monday)
2. Meeting days from C1 (exclude baseline calendar)
3. Available Deep Work slots = (free days) x 2
4. Committed = Deep Work tasks in B5/B6
5. Remaining = Available - Committed

Output: table + recommendation ("Room for N more" or "Overcommitted by N").

STEP 8: Last Week Retro (MONDAY only)

Uses: A4, B3, B5, C3, C1. No new I/O.

1. Shipped: B3 tasks (pre-cleanup) + C3 git logs. Present Repo | Commits | Category table. Flag untracked commits.
2. Wasted: B5 tasks with _notion_edited > 7 days (planned but untouched).
3. Enjoyed: AskUserQuestion — "Energized" / "Mixed" / "Low energy"
4. Meta-KPIs + Spear Sharpening: Read references/meta-kpis.md for formulas.

STEP 9: Task Ranking (MONDAY only)

Uses: B5, B6, B7. No new I/O.

Candidates: all tasks with status Not started, Planned this week, or In progress. Read references/scoring-rules.md for scoring criteria.

Present: Rank | Task | Project | Impact | Energy | Score. Use AskUserQuestion multiSelect for [Owner]'s top 5 (pre-select by score). After confirmation: chosen → "Planned this week", non-chosen → "Not started".

STEP 10: Commit (MONDAY only)

git add notion-mirror/
git commit -m "chore: mission monday — plan week of [date]" </dev/null 2>&1
git push

See _shared/conventions.md for commit-and-verify process.

Behavioral Rules

1. Top-down always. Scorecard → patterns → tasks. Never start with tasks.
2. Three-phase execution. Collect (Step 2) → analyze (Steps 3-9) → interact. No I/O during analysis.
3. Parallel I/O. Groups A, B, C fire simultaneously. Calls within groups also parallel.
4. Pulse is fast. 2-3 minutes, zero interaction unless critically stale.
5. Monday is interactive. AskUserQuestion for energy, KPIs, task selection. Structured 2-4 options.
6. Coaching is separate. /goaly-coaching-prep for [Coach] sessions.
7. See _shared/conventions.md for: tables-not-prose, qmd-first search, commit-and-verify.

Run Log

After presenting output, append a run log entry:

bash ".claude/skills/_shared/scripts/append-run-log.sh" "${CLAUDE_SKILL_DIR}" "mode=MONDAY|PULSE execution=<score> energy=<score> tasks_planned=<count> frog_eaten=<true|false> frog_streak=<count>"

Session Learning

When [Owner] corrects your output or expresses a preference during this session, immediately append it to the feedback log:

mkdir -p "${CLAUDE_SKILL_DIR}/data" && echo "$(date +%Y-%m-%d) <preference description>" >> "${CLAUDE_SKILL_DIR}/data/feedback.log"

Only log general preferences that apply to future sessions — skip task-specific corrections. Use judgment on detail: some preferences are one line, others need a sentence or two of context to be useful in future sessions.

Related Files

Purpose	Path
Mission gotchas	gotchas.md (this directory)
Impact scoring criteria	references/scoring-rules.md
Ephemeral KPI formulas	references/meta-kpis.md
Shared conventions	.claude/skills/_shared/conventions.md
KPI files	notion-mirror/kpis/
Task files	notion-mirror/tasks/
Client files	notion-mirror/clients/
Interaction files	notion-mirror/interactions/
MEMORY.md	~/.claude/projects/-Users-dan-Admin-Goals---Tasks/memory/MEMORY.md
Revenue strategy	~/.claude/projects/-Users-dan-Admin-Goals---Tasks/memory/revenue-strategy.md
Sync log	.sync-log/push.log