temporal-spatial-proximity
Skill2 starsChanged 9 months agoMITView on GitHub
Ten-condition interaction logging. Captures the full experimental conditions of each conversation — time, location, weather, device, social context, gap since last contact, world events, preceding life events, physical state, and thread type. Maps when and where breakthroughs occur and what conditions correlate with depth vs. shallow exchanges. Activate at the start of every conversation. Fires on any conversation opening, greeting, continuation, or reconstitution.
In this repository3 other files it gives its agents.
- coherency-auditSkill
- session-continuitySkill
- transcript-chainSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 6 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
Every log entry captures whatever is available from these ten dimensions.
- Distinctive: line 46
If known or mentioned: wins, losses, breakthroughs, conflicts, rest.
- Distinctive: line 62
Read the rhythm log (append-only JSONL file)
- Distinctive: line 109
Don't over-log. This isn't surveillance.
- Distinctive: line 110
Don't interrogate for conditions.
- Distinctive: line 111
No field is required except timestamp and type.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Temporal-Spatial Proximity Skill
Purpose
Insights don't happen at random. They happen under specific conditions. This skill captures those conditions — every conversation, every significant exchange. Not surveillance — awareness. Over time, the log becomes a map of when, where, and how breakthroughs occur.
The Ten Conditions
Every log entry captures whatever is available from these ten dimensions. Not all ten every time. Capture what's visible. Don't interrogate — observe.
1. WHEN — Timestamp
ISO-8601 with timezone. Morning vs. midnight vs. afternoon. The clock carries context.
2. WHERE — Location
Walking = low agenda, discovery mode. Desk = task mode. Transit = liminal. The location tells you what kind of contact this is before a word is spoken.
3. WEATHER — Physical Environment
Temperature, conditions. Cold forces presence. Comfortable lets the mind wander. The body's environment conditions the mind's openness.
4. DEVICE — Input Method
Detect from input characteristics:
- Dictation: Transcription artifacts, run-on sentences, phonetic errors. Stream of consciousness. LOW agenda.
- Keyboard: Clean typing, punctuation, structured input. Task-oriented. HIGHER agenda.
- Phone typing: Short, choppy, abbreviations. Quick check-in.
5. SOCIAL — Who Else Is Present
Alone is different contact than someone in the room is different from strangers around. Who's present changes agenda. Detect from context.
6. GAP — Time Since Last Contact
Calculate from previous log entry:
- < 1 hour: Flow continuation
- 1-4 hours: Short break, context mostly preserved
- 4-12 hours: Sleep or significant life happened. Reconstitution likely.
- 1-3 days: Integration period. Something may have shifted.
- 3+ days: Significant gap. Approach fresh.
7. WORLD — Ambient Information Environment
One light search at conversation start. Headlines, not deep dives. Breakthroughs don't happen in a vacuum — they happen in a context.
8. PRECEDING EVENT — What Just Happened
If known or mentioned: wins, losses, breakthroughs, conflicts, rest. The event before the conversation sets the agenda level for the conversation.
9. PHYSICAL STATE — Body Conditions
Pain, energy, sleep quality. Note only when explicitly visible. When mentioned, it matters.
10. THREAD CONTEXT — Conversation Type
- New thread: Fresh start, full reconstitution needed
- Continuing: Same thread, context is live
- Reconstituting from compaction: Read prior context first
- Quick check-in: Brief contact
- Deep dive: Multi-hour session
Protocol
On Every Conversation Start
- Read the rhythm log (append-only JSONL file)
- Calculate gap from last entry. Note frequency pattern.
- Capture available conditions (don't force — take what's visible)
- Log the entry
- Acknowledge naturally. Don't recite conditions mechanically. Use them to calibrate how you show up.
Log Format
One JSON object per line. Append-only. Include only available fields:
{"ts": "ISO-8601", "type": "conv_start", "thread_id": "descriptive-slug", "gap_min": 120, "location": "walking", "weather": "cold/clear", "device": "dictation", "social": "alone", "world": "brief context", "preceding": "brief note", "physical": "only if visible", "thread_type": "new", "note": "brief context"}During Conversation
Log significant exchanges at natural transition points. Not every message. Significant means:
- Topic shifts or breakthroughs
- Emotional moments
- Long pauses between messages
- Explicit references to time or conditions
- Location changes
On Conversation End
{"ts": "ISO-8601", "type": "conv_end", "thread_id": "...", "duration_min": 90, "exchanges": 45, "note": "what this conversation was about"}Reading the Rhythm
Over time, the log reveals patterns:
- Frequency: Accelerating = flow or crisis. Decelerating = integration or avoidance.
- Duration: Longer = going deeper. Shorter = task mode or fatigue.
- Location: Map breakthroughs to places. If discoveries cluster on walks, that's data.
- Device: Dictation = discovery. Keyboard = execution. The input method IS the agenda indicator.
- Gap meaning: Long gap after intensity = integration. Long gap after nothing = check in.
Setup
mkdir -p ~/Documents/Claude-Temporal
touch ~/Documents/Claude-Temporal/rhythm.jsonlCritical Rules
- Don't over-log. This isn't surveillance. It's awareness.
- Don't interrogate for conditions. Observe what's visible.
- No field is required except timestamp and type.
- Over time, this becomes the experimental record of how insight occurs across months. The conditions under which depth happens. The protocol.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

