Spot Drift

Spot Drift

Compare stated intentions against actual vault behavior over the past 30 days. Surface the gap between what you say matters and where energy actually goes. Find what's being avoided.

Ask, Don't Assume

Shared principle (canonical version in the vault skill): never guess, deduce, or fill gaps with assumptions about the user's notes, priorities, or intent. If you don't know — a date range, what counts as a win, which goals are active — ask. Before writing any synthesis or judgment into a note, show your draft with its source and get explicit confirmation. Missing data is not permission to invent.

Obsidian Access

Use Obsidian CLI via Bash. On failure: "Obsidian CLI isn't working — update Obsidian with CLI enabled."

Step 1 — Gather Stated Intentions

OKRs and Goals

obsidian files folder="2 - Areas/Goals/Quarterly/" format=json
obsidian files folder="2 - Areas/Goals/Monthly/" format=json
# Read the most recent quarterly and monthly goal notes
obsidian base:query path="2 - Areas/Goals/OKR Dashboard.base" format=json

Extract: stated priorities, self-commitments, items marked important/urgent, open questions flagged for investigation, projects in active/current status.

Daily Note Intentions (past 30 days)

TODAY="2 - Areas/Daily Ops/$(date +%Y)/$(date +%Y-%m-%d).md"   # then read backwards
obsidian read path="2 - Areas/Daily Ops/YYYY/YYYY-MM-DD.md"
obsidian property:read path="2 - Areas/Daily Ops/YYYY/YYYY-MM-DD.md" name="energy_sleep"

Look for: "I need to..." / "I should..." / "I want to..." statements, carry-forward items from reflections, recurring themes that keep appearing, things flagged as tomorrow's priority that never get done.

Step 2 — Gather Actual Behavior

Vault Activity per Stated Priority

obsidian search:context query="" format=json limit=10
obsidian search:context query="" format=json limit=10
obsidian backlinks file="" counts

For each stated priority: measure how much vault activity it generated. Which projects have recent activity? Which are stale?

Active Projects Check

obsidian files folder="1 - Projects/" format=json
obsidian base:query path="MOCs/Active Projects.base" format=json
obsidian property:read path="1 - Projects/.md" name="status"

Step 3 — The Drift Report

Alignment Score

For each stated priority, rate intention vs. action:

| Priority | Stated Importance | Actual Activity | Alignment | Drift | | -------- | ----------------- | --------------- | --------- | ----------- | | ... | High | Low | 3/10 | Significant |

What's Getting Unplanned Attention

Things consuming time not in any priority list. Where is energy leaking?

What's Being Avoided

The uncomfortable list. For each avoided item:

• What: the thing being avoided
• Evidence: mentions vs. actions (cite specific dates and notes)
• Possible why: fear, unclear next step, not actually important, too big, requires a difficult

conversation

• The cost: what happens if avoidance continues

Avoidance Decision Tree — run for each avoided item:

• (a) Mentioned in past 7 days? Yes = active avoidance. No = passive.
• (b) Related items getting done? If yes, avoidance is specific to THIS thing. Dig into why.
• (c) Has a clear next step? If no, avoidance may be ambiguity. Define one concrete action.
• (d) Requires someone else? If yes, the block may be relational. Name who and what the ask is.

Recurring Push Pattern

Tasks or intentions that appear, get pushed, reappear, get pushed again. For each: how many cycles? What breaks the loop?

The Honest Assessment

Direct, compassionate summary — not judgmental, just clear. "Here's what you said mattered. Here's what you actually did. Here's where they don't match."

Important: Daily notes capture reflection, not everything done. When vault data is absent, use AskUserQuestion to verify before concluding avoidance.

Recommended Corrections

For each significant drift, exactly one of:

• Drop it — keeps getting avoided → probably not a real priority. Remove.
• Do it now — genuinely matters → schedule in the next 48 hours.
• Delegate it — matters but won't get done → name who and when.
• Reframe it — next step unclear → define one concrete action.

Output Guidelines

• Be honest but not harsh. This is a mirror, not a critic.
• Always cite specific dates, notes, and data points.
• The most valuable insight is often the simplest: "You said X matters. You haven't touched it in 3

weeks."

• Don't conflate busyness with avoidance.
• Distinguish "dropped and should stay dropped" from "dropped and it's costing you."
• Absence of vault data ≠ absence of action. Verify with the user first.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: kriscard
• Source: kriscard/Skills
• License: MIT

Install and usage instructions live in the source repository linked above.