flashlight
Shine a light into a wayfinder map's fog — work one direction now, out of frontier order, or redraw the map itself.
Skill by will-ness-ai
Chalk Labs
Need help deploying flashlight?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/will-ness-ai-flashlight
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/flashlight
What's inside
- Skill name
flashlight- Description
- Shine a light into a wayfinder map's fog — work one direction now, out of frontier order, or redraw the map itself.
- License
- MIT
- Instructions
- 101 lines in SKILL.md · view on GitHub
Files (2)
- SKILL.md
- openai.yaml · agents/
SKILL.md (first lines)
A **wayfinder map** charts a large effort as decision tickets on the issue tracker, wrapped in **fog**. Its **frontier** — the open, unblocked, unclaimed tickets — advances in its own order. A flashlight is the human's beam. They point it at one direction, and this session walks in **now**, out of frontier order, to learn what lies that way. `/flashlight <map> <request>`. Ask for the map when the invocation omits it. Shine **one beam per session**. Where the map, its child tickets, blocking edges, claims and the frontier query physically live is tracker-specific. Consult the tracker doc's "Wayfinding operations" section for how this repo expresses each one. Default to the local-markdown tracker when no tracker doc is present. ## Two beams Read `<request>`, and ask one question: **is the subject the map, or the territory?** - **Territory** — the work itself. "Can we use SQLite for the queue?" → [Shine into the territory](#shine-into-the-territory). - **Map** — the chart you already drew: what to focus on, what to defer, what no longer belongs. "We have too many open tickets — what should we focus on?" → [Redraw the map](#redraw-the-map). Ask the human to choose when the request reads both ways. ## Shine into the territory The beam becomes one ad-hoc ticket, and leaves the same trail as any ticket the map was charted with. ### 1. Aim the beam Load the map — Destination, Notes, Decisions so far — and the open tickets bearing on the direction. Where an open ticket already asks this question, work that one instead: claim it, and continue from step 4. Infer the **type** from the request: | Type | Use it for | Resolve with | | --- | --- | --- | | **Grilling** (HITL) | a decision to settle by conversation — the default | `/grilling` and `/domain-modeling` | | **Research** (AFK) | a fact from outside this working directory | a `/research` subagent | | **Prototype** (HITL) | "how should it look, how should it behave", answered by a cheap concrete artifact | `/prototype`; `/grill-design` for an unsettled frontend look | | **Task** (HITL or AFK) | manual work that unblocks a decision | the agent alone where it can; otherwise a precise checklist for the human | On a **HITL** type the human answers their own questions — this session asks and records. An **AFK** type runs alone. **Done when** the type is chosen and the map's Destination is in hand. ### 2. Zoom out Brief the human: the **type** in bold, what this session is about to do, and where the direction sits on the map. **Done when** the human has seen the type before any ticket exists. ### 3. Create and claim the ticket Create a child issue of the map titled `Flashlight: <direction>`, carrying the verbatim `<request>` as its question and the `wayfinder:<type>` label. It takes no blocking edges — it is born and closed inside this session. **Claim it** — assign it to the dev driving the map — before the work starts. **Done when** the ticket exists, carries its type, and shows an assignee. ### 4. Work the ticket Resolve it as a ticket of its type. Invoke the skills the map's `## Notes` names, plus the type's own. Zoom into any related or closed ticket on demand. **Done when** the question has an answer its type can support. ### 5. Record and adjust the map - Post the answer as a **resolution comment**, then **close** the ticket. - Append one line to **Decisions so far**: the ticket name, its link, and a one-line gist of the answer. - Create the tickets the answer made specifiable, and clear each graduated patch from **Not yet specified**. - Rule out of scope whatever the answer put past the destination: close it, and leave one line in **Out of scope**. **Done when** the ticket is closed and every section the answer touched carries its line. ## Redraw the map The beam falls on the chart itself. This mode creates no ticket — the comments it leaves on the tickets it touches are the record. ### 1. Read the whole chart Load the map and **every** open ticket: question, type, blockers, assignee. **Done when** each open ticket is accounted for, the claimed ones included. ### 2. Propose Show the human one change list, and wait for their approval. Name every edit, and why it earns its place. **Done when** the human has approved the list. ### 3. Apply | Verb | What it does | | --- | --- | | **Create** | a new ticket — then wire its edges in a second pass, because issues need ids before they can reference each other | | **Update** | a ticket's question | | **Re-order** | map order decides frontier order, so this is how you prioritise | | **Re-wire** | a ticket blocked by an open ticket leaves the frontier and returns when that blocker closes, so this is how you defer | | **Rule out of scope** | close it, and leave one line in **Out of scope**: the gist, and why it sits past the destination | | **Cancel** | for a duplicate or a mistake. Close it, and leave one line in **Decisions so far** marked **cancelled**, carrying the reason | Comment on every ticket you touch, saying what happened and why. A ticket with an assignee is **claimed**: a live session is working it right now. Leave its state as it is, and comment on it, so that session learns what changed. **Done when** every approved edit is applied, and every touched ticket carries its comment.
From will-ness-ai/skills, MIT.
More from will-ness-ai/skills
- 0.786score
agent-tools
will-ness-ai
Build commands for agents in AXI style.
SkillAFreshness: A★ 170Claude Code +9 - 0.742score
sandcastle
will-ness-ai
Run a Sandcastle lane — a sandboxed implement→review loop over a fixed batch of work in its own worktree, ended by an agent-authored PR.
SkillBFreshness: B★ 170Claude Code +9 - 0.739score
code-story
will-ness-ai
Build a wizard-style HTML page that teaches how and why a change works.
SkillBFreshness: B★ 170Claude Code +9 - 0.739score
cmux
will-ness-ai
Drive the local cmux app — workspaces, panes, surfaces, terminal input, agent sessions, and browser surfaces.
SkillBFreshness: B★ 170Claude Code +9 - 0.734score
grill-design
will-ness-ai
Converge on a frontend look through rounds of prototypes and grilling verdicts.
SkillBFreshness: B★ 170Claude Code +9 - 0.607score
test-a-skill
will-ness-ai
Field-test a skill by running it cold in parallel agent sessions, then turn what they hit into edits.
SkillBFreshness: B★ 170Claude Code +9 - 0.352score
find-standards
will-ness-ai
No description from the publisher yet.
SkillBFreshness: B★ 170Claude Code +9 - 0.349score
grilling-frontend-prototyping
will-ness-ai
No description from the publisher yet.
SkillBFreshness: B★ 170Claude Code +9
Signals
- Stars
- 170
- Forks
- 4
- Open issues
- 0
- Last commit (this folder)
- Aug 11, 2026 (2 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/will-ness-ai/skills
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from will-ness-ai
- 0.786score
agent-tools
will-ness-ai
Build commands for agents in AXI style.
SkillAFreshness: A★ 170Claude Code +9 - 0.742score
sandcastle
will-ness-ai
Run a Sandcastle lane — a sandboxed implement→review loop over a fixed batch of work in its own worktree, ended by an agent-authored PR.
SkillBFreshness: B★ 170Claude Code +9 - 0.739score
cmux
will-ness-ai
Drive the local cmux app — workspaces, panes, surfaces, terminal input, agent sessions, and browser surfaces.
SkillBFreshness: B★ 170Claude Code +9 - 0.739score
code-story
will-ness-ai
Build a wizard-style HTML page that teaches how and why a change works.
SkillBFreshness: B★ 170Claude Code +9 - 0.734score
grill-design
will-ness-ai
Converge on a frontend look through rounds of prototypes and grilling verdicts.
SkillBFreshness: B★ 170Claude Code +9 - 0.607score
test-a-skill
will-ness-ai
Field-test a skill by running it cold in parallel agent sessions, then turn what they hit into edits.
SkillBFreshness: B★ 170Claude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get flashlight running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
flashlight skill. Agent Mart. https://agentmart.dev/skills/will-ness-ai-flashlight (last updated 2026-10-24)