brutal-wayfinder
Skill2 starsChanged 4 months agoMITView on GitHub
Map and resolve planning work whose route is still hidden by research, prototypes, access, or decisions spanning multiple sessions. Use for large or foggy efforts that cannot yet produce a decision-complete Brutal specification and implementation tickets.
In this repository25 other files it gives its agents.
- brutal-deepresearchSkill
- brutal-grillSkill
- brutal-idea-evalSkill
- brutal-inf-fix-loopSkill
- brutal-planSkill
- brutal-pr-finding-fixerSkill
- brutal-pr-fix-loopSkill
- brutal-project-reviewSkill
- brutal-pr-reviewSkill
- brutal-realitySkill
- brutal-reviewSkill
- brutal-specSkill
- brutal-swarmSkill
- brutal-ticketsSkill
- brutal-workerSkill
- codex-coopSkill
- design-jepa-modelsSkill
- factory-designSkill
- funcvec-dedupe-fixerSkill
- grill-meSkill
- handoffSkill
- opencode-delegateSkill
- rust-architecture-reviewSkill
- rust-pr-reviewSkill
- supply-chain-analyzerSkill
The file
We split this skill into the rules it gives an agent
10 rules. About 0 are common, 4 are distinctive, and 6 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
6
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 14
Read ../brutal-shared/integration-resolver.md, resolve the work store, and load its support module.
- Distinctive: line 17
Read repository instructions, domain context, ADRs, and existing map state.
- Very specific: line 18
Use type:investigation for every map and investigation ticket.
- Very specific: line 23
Use Source: brutal-wayfinder.
- Very specific: line 51
Use brutal-grill to name and confirm the destination.
- Very specific: line 56
If no fog remains and the decisions fit the current planning conversation, return to brutal-plan without creating a map.
- Distinctive: line 61
Create the map first, then child tickets, then native relationships where supported.
- Distinctive: line 73
Ask whether this run should resolve one frontier question or continue through ready agent-only questions.
- Very specific: line 79
Add newly visible questions, graduate sharpened fog into tickets, and move newly excluded work to Out Of Scope.
- Distinctive: line 82
Use adapter assignment, states, comments, parents, and blockers.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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
Brutal Wayfinder
Persist a resumable map of questions until the route to a decision-complete specification is clear. Produce planning decisions, not destination implementation.
Required Context
- Read
../brutal-shared/integration-resolver.md, resolve the work store, and load its support module. - Read
../brutal-grill/SKILL.md. - Read repository instructions, domain context, ADRs, and existing map state.
- Use
type:investigationfor every map and investigation ticket. These artifacts never enter the implementation task queue.
Map Contract
Use Source: brutal-wayfinder. A map contains:
## Destination
## Notes
## Decisions So Far
## Not Yet Specified
## Out Of ScopeThe destination states what must exist when wayfinding ends. Decisions So Far
is an index of resolved ticket names and one-line outcomes, not a duplicate of
their full resolutions. Not Yet Specified holds in-scope fog that cannot yet
be phrased as a precise question.
Each child ticket contains its parent, one precise Question, blockers, claim,
status, and one mode:
research— agent-only primary-source investigationprototype— human-in-the-loop artifact for behavior or design feedbackgrilling— human-in-the-loop decision usingbrutal-grilltask— prerequisite work needed to expose facts for a later decision
A prerequisite task must unblock a planning decision; destination implementation belongs in the later Brutal plan.
Chart A Map
- Use
brutal-grillto name and confirm the destination. - Map breadth-first: group questions that must be researched or decided together into one coherent investigation scope. Create separate tickets only for different blockers, modes, owners, or independently useful decisions; record the remaining fog without prematurely slicing it.
- If no fog remains and the decisions fit the current planning conversation,
return to
brutal-planwithout creating a map. - Present the destination and a one-line list of initial scopes, blockers, and fog. Publish only after explicit approval; show full ticket bodies only when requested.
- Create the map first, then child tickets, then native relationships where supported. Mirror parent and blocker references in every ticket body.
Persist the map and children through the resolved work-store contract. Use the local adapter's investigation paths or remote parent/child artifacts, logical queue states, and prepare-before-create strategy. Never create local work-store artifacts for a remote adapter.
Work A Map
- Load the map as the low-resolution index and query open children.
- Identify the frontier: open, unblocked, unclaimed tickets.
- Ask whether this run should resolve one frontier question or continue through ready agent-only questions.
- Claim a ticket before work. In continuous mode, pause before any human-in-the-loop ticket.
- Record the full resolution on the ticket, mark it done, and append its linked
one-line outcome to
Decisions So Far. - Add newly visible questions, graduate sharpened fog into tickets, and move
newly excluded work to
Out Of Scope.
Use adapter assignment, states, comments, parents, and blockers; preserve the same information in bodies when the provider has no native representation.
Complete the map only when no open ticket or fog remains. Mark the parent done
and hand its reference to brutal-plan, which consumes the decisions, proposes
any domain-document changes, and links the map from the final specification.
On rerun, deduplicate by Source, parent reference, and exact ticket title.
Report partial publication or update failures precisely so the next run can
resume rather than duplicate artifacts.
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.

