wayfinding
Skill2 starsChanged 49 days agoMITView on GitHub
Charts work too big or too foggy for one session as a map of decision tickets in project memory, then resolves one per session until the way is clear. Use when an idea, a product or a change is too large for one sitting, or to continue a map.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- tslateman/skillstheir own version, first seen Oct 11, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- tslateman/skills★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository25 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- agent-designSkill
- agents-mdSkill
- bug-diagnosisSkill
- builder-osSkill
- business-caseSkill
- delivery-disciplineSkill
- domain-modelSkill
- evidence-ledgerSkill
- gate-checksSkill
- go-to-marketSkill
- ideation-methodsSkill
- opportunity-mappingSkill
- outcome-reviewSkill
- pressure-testingSkill
- problem-framingSkill
- product-analysisSkill
- release-opsSkill
- research-methodsSkill
- spec-writingSkill
- system-designSkill
- tracking-standardsSkill
- using-builder-osSkill
- ux-architectureSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 60
Every ticket is worked with the person (they speak for themselves) or alone by the agent.
- Very specific: line 97
Record it: Write ## Resolution, set Status: closed, add one line under Decisions so far.
≈ “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
Wayfinding
A loose idea has arrived, too big for one session and wrapped in fog: the way from here to the destination is not visible yet. Wayfinding finds the way; it does not charge at the destination. It writes a map into project memory and works its decision tickets (questions whose answer is a decision, not slices of a build) one per session, until nothing is left to decide before someone does the work.
Pocock's wayfinder (MIT) is the source of the method; this version keeps the map in .builderos/ so it works with nothing connected and every new session reads it.
Plan, don't do
Each ticket resolves a decision; the map is done when the way is clear. The pull to start building is the signal you have reached the edge of the map: hand off. A map can carry execution only when its Notes say so.
Where the map lives
.builderos/maps/{slug}/MAP.md is the index, one file per ticket in tickets/NN-{slug}.md, numbered in creation order. The map is an index, not a store: a decision lives in its ticket, the map gists it and links. Refer to every ticket by its title, with the number inside the link, never by a bare number. Where tickets.write resolved and the person wants the tracker, publish each ticket there too, with native blocking edges, and write the link into its file; the files stay the record.
Where commands run, scripts/bos.mjs map prints the destination, the frontier, what is claimed, blocked and decided, and the fog; the session briefing shows the next takeable ticket of every open map.
# Map — {name}
**Status:** open
## Destination
{What reaching the end looks like: initiatives on the roadmap with their tracks, a spec ready for phase 4, a decision recorded as an ADR, a change made. One or two lines.}
## Notes
{Domain, skills every session reads, standing preferences for this effort.}
## Decisions so far
- [{ticket title} (01)](tickets/01-{slug}.md): {one-line gist of the answer}
## Not yet specified
- {A dim question you can see coming but cannot phrase sharply yet}
## Out of scope
- {Work ruled beyond the destination, with why, linking the closed ticket if there was one}# 01: {question as a title}
**Type:** research | prototype | grilling | task
**Status:** open | claimed | closed | out-of-scope
**Blocked by:** {numbers of the tickets that gate this one, or none}
**Claimed by:** {who, date; empty while unclaimed}
## Question
{The decision or investigation, sized to one session.}
## Resolution
{Written on close: the answer, who decided it, the evidence tags, links to what was produced.}A ticket is unblocked when every ticket blocking it is closed. The frontier is the open, unblocked, unclaimed tickets: the edge of the known.
Ticket types
Every ticket is worked with the person (they speak for themselves) or alone by the agent. A ticket worked with the person resolves only through that exchange: the person's answer closes it.
- Research (alone): a fact a decision waits on, from documents, the repository, the web or connected sources. Facts are the agent's job. Delegate it where the host supports delegation, so research tickets resolve in parallel; otherwise resolve it inline. Tag what it finds per
evidence-ledger. - Prototype (with the person): raise the fidelity of the discussion with a cheap, rough artifact to react to: a logic prototype per
domain-model, rendered versions perux-architecture(direction), an outline. Link it from the ticket. - Grilling (with the person): the default. Run
pressure-testingrounds until the ticket's frontier is empty; settle any term used two ways perdomain-model, step 1. - Task (alone where possible, otherwise a checklist for the person): manual work a decision is blocked on: an account to open so an API can be judged, an export so data can be seen. It earns its place by unblocking a decision. Its resolution records what was done and the facts later tickets need.
Fog of war
The map is deliberately incomplete. Not yet specified holds the dim view: decisions you can tell are coming but cannot phrase yet, because they hang on open questions. Resolving a ticket clears fog ahead of it.
Fog or ticket? Ticket when you can state the question precisely now, even if it is blocked. Fog when you cannot phrase it that sharply yet; one patch may become several tickets, or none.
Fog gathers only toward the destination. Work beyond it is out of scope: close any ticket that turns out to sit past the destination, with status out-of-scope, and leave one line under Out of scope. It never graduates; it returns only if the destination is redrawn, as a new map.
When a closed decision turns out wrong
A resolution can fail later: a fact proves false, a prototype contradicts it, the person changes their mind. Reopen it rather than working around it: set the ticket back to **Status:** open, clear **Claimed by:**, keep the old resolution under ## Superseded with the date and why, and strike its line in Decisions so far. Every closed ticket that lists it under Blocked by, directly or through another, is reopened the same way, because its answer rested on the one now gone. Done when no closed ticket rests on a reopened one, and bos.mjs map shows them back on the frontier.
Procedure
One ticket per session, research tickets excepted: the next session starts from the redrawn map, not from this one's momentum.
Chart the map
- Name the destination. One
pressure-testinground on what this map is finding its way to; the destination fixes the scope, so it is settled first. Done when the person accepted a destination of one or two lines. - Map the frontier, breadth first. Fan out across the whole space, not deep on one thread: the open decisions and the first steps takeable now. If the way is already clear and fits one session, say so and offer the direct route instead (an initiative, a spec, an answer). Keep the destination short of a waterfall: when the first frontier holds more than about six tickets, split it into maps run one after another, the first ending where something can be shown; and when the destination has a look or a behavior to react to, put a prototype ticket on the first frontier. Done when every open decision you can see is either a sharp question or a patch of fog.
- Write the map with Destination and Notes, an empty Decisions so far, and the fog under Not yet specified. Done when
MAP.mdexists under.builderos/maps/{slug}/with all five headings. - Write the tickets you can specify now, then wire the blocking edges in a second pass, with real numbers. Done when every ticket has a type, a question and its blockers.
- Start the research tickets, in parallel where delegation is available. Done when each research ticket is claimed or resolved.
- Stop. Charting is one session's work. Done when the person has the destination, the frontier and the fog, and no grilling ticket has been worked.
Work through the map
- Load the map, not every ticket body (
scripts/bos.mjs map {slug}where commands run). Done when you can name the frontier, the claimed tickets and the fog. - Choose the ticket. The one the person named, otherwise the first on the frontier. Claim it first (
**Status:** claimed,**Claimed by:**), so a parallel session skips it. Done when the claim is written in the ticket file, re-read just before writing. - Resolve it as its type says, reading any related or closed ticket on demand and the skills the Notes name. Done when the ticket's own question has an answer the person gave (with them) or a sourced fact (alone).
- Record it. Write
## Resolution, set**Status:** closed, add one line under Decisions so far. A decision that is hard to reverse, surprising and a real trade-off also becomes an ADR indecisions/. Done when the ticket, the map and any ADR all carry the decision. - Redraw. New tickets the answer surfaced (write, then wire); fog the answer made specifiable graduates into tickets and leaves Not yet specified; tickets the answer invalidated are updated, or closed as out of scope. Done when the map matches what is now known.
- When no ticket and no fog remain, the way is clear: set
**Status:** doneand hand off to the destination:builder-osfor initiatives (each with its track and the map linked as its evidence),spec-writingfor a spec, the ADR for a decision. Done whenbos.mjs mapshows no open ticket and Not yet specified is empty.
Expect other sessions to be working the same map: re-read a ticket before claiming it.
Capabilities
| Capability | Used for | Floor if absent |
files.read / files.write | The map and its tickets | Keep the map in the conversation and say it will not survive the session |
subagent.dispatch | Research tickets in parallel | Resolve them inline, one after another |
tickets.write | Publishing tickets to the person's tracker | The files are the tracker |
shell.exec | bos.mjs map for the frontier | Read the ticket files and compute it by hand |
Output Contract
Charting ends with ## MAP CHARTED: the destination, the frontier by title, the fog. A worked ticket ends with ## TICKET RESOLVED: the decision, who made it, what it unblocked, the next ticket on the frontier. A finished map ends with ## WAY CLEAR and the hand-off.
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.

