The Time Capsule Skill
Handover docs describe systems; nobody writes down the thing successors
actually need — what you believed and how sure you were. A time capsule is a
memo with a seal date and an open date: the honest state of things, named
predictions with confidence levels, the advice you'd whisper across time — and
a scoring ritual, because a prediction you never score is just a mood. It's
calibration infrastructure wearing an emotionally sticky costume, and it works
on future-you exactly as well as on a successor.
What This Skill Produces
- The capsule memo: honest state, the things-nobody-writes-down (fragile
truces, bodies buried, bets in flight), and advice offered with humility
- A prediction ledger: 5–10 falsifiable predictions, each with a confidence
percentage, a resolution date, and its scoring criterion
- The seal: open-on date and trigger ("when you're considering rewriting
the billing system, open early")
- An opening-day ritual: how to score the ledger and what to do with what
it reveals — feeds [[decision-journal]] and, at 90 days, dovetails with
[[manager-first-90-days]] for role handovers
Required Inputs
Ask for (if not already provided):
- Who opens it (successor, future you, the team) and roughly when (a date, or
a trigger event)
- The occasion: leaving, project end, year-end, pre-leave
- The current state as they'd tell a trusted friend — including what's fragile,
what's pretending to be fine, and what they'd do next if they stayed
- The calls they're least sure about (those become the best predictions)
Process
- Write the state honestly. Three registers: what's working (and why,
which is what actually breaks), what's fragile (the truce with team X, the
vendor coasting on goodwill), what's pretending (the metric everyone quotes
that no longer means anything). The rule: nothing in the capsule the writer
wouldn't want quoted back — it will be, on opening day, by them.
- Extract falsifiable predictions. Convert beliefs to scoreable claims:
"the migration lands by Q3" → "Migration in production by Sept 30 —
70%." Each gets: claim, confidence %, resolution date, and what counts as
true (settle the scoring argument now, not later). Push for at least two
predictions the writer is uncomfortable writing — those carry the
calibration signal.
- Give advice as bets, not commandments. "If X happens, I'd do Y, because
Z" ages well; "never reorganize the platform team" ages into a dare.
Include one "permission slip" — the thing the successor will hesitate to
change that the writer hereby blesses changing.
- Seal it properly. Open-on date + early-open triggers + where it lives
(a file with the date in its name; calendar reminder for the opener). Note
what the writer commits to NOT doing: editing it after sealing.
- Script opening day. Score each prediction right/wrong/unresolvable ·
compute the honest hit rate vs stated confidence · one paragraph: "what
would past-me be surprised by?" · log the misses worth learning from in
[[decision-journal]]. Then — the ritual's point — write the next capsule.
Output Format
# 🕰 Time capsule — sealed [date], open [date/trigger]
To: [opener] · From: [writer, role, occasion]
## The honest state
[Working & why · fragile · pretending]
## Prediction ledger
| # | Claim (falsifiable) | Confidence | Resolves | Counts as true if |
## Advice, offered as bets
[If-then-because lines · the permission slip]
## What I'd do next if I were staying
[The plan they never got to run]
---
## Opening-day ritual (don't read past the seal until then)
[Score the ledger · hit-rate vs confidence · surprises paragraph · log misses
→ decision-journal · write the next capsule]
Quality Checks
Anti-Patterns
Example Trigger Phrases
- "Leave a role."
- "Finishing a big project."
- "Plan season."
- "Before a leave."
- "Write a letter to my successor."
1---2name: the-time-capsule3description: "Write a sealed memo to your future self or successor — the honest state of things, falsifiable predictions with confidence levels, and the advice you suspect they'll need — with an open-on date and a scoring ritual for when it's opened. Use when leaving a role, finishing a big project, at year-end or planning season, before a leave, or 'write a letter to my successor'. Produces the sealed capsule, its prediction ledger, and the opening-day ritual."4disable-model-invocation: true # skill-optimizer5metadata:6 source: https://skillmd.com/skills/slipcast-dev/the-time-capsule7---89# The Time Capsule Skill1011Handover docs describe systems; nobody writes down the thing successors12actually need — *what you believed and how sure you were*. A time capsule is a13memo with a seal date and an open date: the honest state of things, named14predictions with confidence levels, the advice you'd whisper across time — and15a scoring ritual, because a prediction you never score is just a mood. It's16calibration infrastructure wearing an emotionally sticky costume, and it works17on future-you exactly as well as on a successor.1819## What This Skill Produces2021- The **capsule memo**: honest state, the things-nobody-writes-down (fragile22 truces, bodies buried, bets in flight), and advice offered with humility23- A **prediction ledger**: 5–10 falsifiable predictions, each with a confidence24 percentage, a resolution date, and its scoring criterion25- The **seal**: open-on date and trigger ("when you're considering rewriting26 the billing system, open early")27- An **opening-day ritual**: how to score the ledger and what to do with what28 it reveals — feeds [[decision-journal]] and, at 90 days, dovetails with29 [[manager-first-90-days]] for role handovers3031## Required Inputs3233Ask for (if not already provided):34- Who opens it (successor, future you, the team) and roughly when (a date, or35 a trigger event)36- The occasion: leaving, project end, year-end, pre-leave37- The current state as they'd tell a trusted friend — including what's fragile,38 what's pretending to be fine, and what they'd do next if they stayed39- The calls they're least sure about (those become the best predictions)4041## Process42431. **Write the state honestly.** Three registers: what's working (and *why*,44 which is what actually breaks), what's fragile (the truce with team X, the45 vendor coasting on goodwill), what's pretending (the metric everyone quotes46 that no longer means anything). The rule: nothing in the capsule the writer47 wouldn't want quoted back — it will be, on opening day, by them.482. **Extract falsifiable predictions.** Convert beliefs to scoreable claims:49 "the migration lands by Q3" → *"Migration in production by Sept 30 —50 70%."* Each gets: claim, confidence %, resolution date, and what counts as51 true (settle the scoring argument now, not later). Push for at least two52 predictions the writer is *uncomfortable* writing — those carry the53 calibration signal.543. **Give advice as bets, not commandments.** "If X happens, I'd do Y, because55 Z" ages well; "never reorganize the platform team" ages into a dare.56 Include one "permission slip" — the thing the successor will hesitate to57 change that the writer hereby blesses changing.584. **Seal it properly.** Open-on date + early-open triggers + where it lives59 (a file with the date in its name; calendar reminder for the opener). Note60 what the writer commits to NOT doing: editing it after sealing.615. **Script opening day.** Score each prediction right/wrong/unresolvable ·62 compute the honest hit rate vs stated confidence · one paragraph: "what63 would past-me be surprised by?" · log the misses worth learning from in64 [[decision-journal]]. Then — the ritual's point — write the next capsule.6566## Output Format6768```69# 🕰 Time capsule — sealed [date], open [date/trigger]70To: [opener] · From: [writer, role, occasion]7172## The honest state73[Working & why · fragile · pretending]7475## Prediction ledger76| # | Claim (falsifiable) | Confidence | Resolves | Counts as true if |7778## Advice, offered as bets79[If-then-because lines · the permission slip]8081## What I'd do next if I were staying82[The plan they never got to run]8384---85## Opening-day ritual (don't read past the seal until then)86[Score the ledger · hit-rate vs confidence · surprises paragraph · log misses87 → decision-journal · write the next capsule]88```8990## Quality Checks9192- [ ] Every prediction is falsifiable with its scoring criterion pre-agreed —93 no "things will improve"94- [ ] Confidence percentages vary; a ledger of all-70% is hedging in costume95- [ ] At least one fragile-or-pretending item that would never appear in an96 official handover97- [ ] Advice is in if-then-because form; the permission slip exists98- [ ] The seal has both a date and an early-open trigger, and the memo commits99 the writer to not editing after sealing100101## Anti-Patterns102103- [ ] Do not write a legacy-polishing document — the opener scores it; vanity104 ages worst of all105- [ ] Do not let predictions hide in prose; the ledger table is the contract106- [ ] Do not settle scores or name-blame — "the truce with X is fragile" is107 information; "X is impossible" is a grenade with a delay fuse108- [ ] Do not skip the ritual section; an unopened capsule is a diary, and an109 unscored ledger is astrology110111## Example Trigger Phrases112113- "Leave a role."114- "Finishing a big project."115- "Plan season."116- "Before a leave."117- "Write a letter to my successor."