Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add OutlineDriven/odin-claude-plugin --skill minimapgit clone --depth 1 https://github.com/OutlineDriven/odin-claude-pluginWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/outlinedriven/odin-claude-plugin/minimap)<a href="https://agentmods.dev/skills/outlinedriven/odin-claude-plugin/minimap"><img src="https://agentmods.dev/badge/skills/outlinedriven/odin-claude-plugin/minimap.svg" alt="Measured on agentmods" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00056 | $0.01367 |
| Opus 5 | $0.00028 | $0.00683 |
| Sonnet 5 | $0.00011 | $0.00273 |
| Haiku 4.5 | $0.00006 | $0.00137 |
Grade A, and why
minimap scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 46 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Minimap
Contract
| Field | Bound contract |
|---|---|
| Trigger | Use on an explicit model or human request for a compact view of current work state, including at an SDD phase boundary. |
| Authority | Read-only: inspect local Git state, recorded test evidence, and root graph.yaml only; never mutate files, VCS, credentials, paid services, publications, deployments, or remote state. Never run tests. |
| Side effect | Emit one compact view and write nothing. |
| Done | Every reported claim has exactly one class (conquered, disputed, or fog) and a concise justification from the evidence join, with absent graph knowledge represented as fog. |
Inputs
- A repository or worktree root, inferred from the current directory when unambiguous.
- Its readable current Git commit, tree identity, tracked changes, conflicts, and relevant branch or upstream state.
- The latest repository test result already recorded by the repository or supplied by the caller. A usable result identifies its checks and outcomes and binds them to a commit identifier or tree digest. Its absence is allowed and produces fog; this skill never creates or refreshes it.
- Optional project-root
graph.yaml. Its absence is allowed and produces fog for graph knowledge while Git-and-test claims are still rendered.
Procedure
- Resolve one repository root and snapshot its read-only Git evidence: current commit and tree identity, tracked dirty paths, conflicts, and branch or upstream facts relevant to the requested view. Do not inspect or report unrelated repositories. Done when: the root is identified and the Git snapshot contains commit SHA, tree identity, dirty path list, conflict list, and branch/upstream state, each read without mutation.
- Locate the latest already-recorded repository test result using its explicit completion time or repository-defined ordering. If no unique latest result can be established, treat applicable test knowledge as missing and record that reason; never run a test or invent an ordering. Done when: one test result is selected as the latest, or no unique latest exists and the reason is recorded so that test knowledge will be fog.
- Accept a recorded test result only when it names its checks and outcomes and carries a commit identifier or tree digest. Compare that binding with the current commit or tree. A matching binding is current; a well-formed nonmatching binding is stale and remains contradiction evidence, not current proof. Done when: the selected result is classified as current or stale against the current commit or tree, or it is rejected as malformed and the rejection reason is recorded.
- Read
references/graph-schema.md. If rootgraph.yamlexists, read it without modification. Require schemaodin.graph/v1, a non-empty revision, unique non-empty node IDs, non-empty labels, and edges whose endpoints exist. Accept only node typesInvariant,Surface,Store,Hazard, andDecision, and edge typesconstrains,reads,writes, andfails-when. Reject an unknown schema version, node type, edge type, duplicate node ID, or dangling edge. Ignore unknown top-level and per-record fields. Treat nodes and edges as immutable assertions within their declared revision; do not infer, replan, or alter topology. Done when:graph.yamlis parsed and every node and edge passes the schema checks, or it is rejected with the specific violation named, or it is absent and graph knowledge is marked fog. - Form the smallest complete claim set from current Git facts, each recorded check outcome, and each valid graph node or edge relevant to the requested work. Preserve source identity for every fact. When
graph.yamlis absent, add a graph-knowledge claim rather than omitting that source. Do not invent claims or applicability links unsupported by the inputs. Done when: every claim in the set traces to a named source (Git fact, test outcome, or graph node/edge), and no claim exists that lacks a source. - For each claim, join all applicable evidence without granting any source global precedence. Classify it
disputedwhen applicable sources contradict, proof is stale, or dirty state affects the paths or tree covered by the claim. Otherwise classify itfogwhen applicable evidence is missing. Otherwise classify itconqueredonly when all applicable evidence is current and agrees. Done when: every claim has exactly one classification, and the classification is justified by the specific evidence sources that were joined. - Render one compact view with exactly the three class headings
conquered,disputed, andfog, including empty headings when a class has no claims. Place every claim under exactly one heading with its supporting source identities and one concise join justification. This single view is the complete SDD-minimap result; emit no second campaign or status map. Done when: the view contains all three headings, every claim appears under exactly one, and each claim line carries its source identities and join justification.
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 4d ago First seen · 46 lines · 56 tokens per session scan A ebfa0e043b60
minimap is a skill published in the GitHub repository OutlineDriven/odin-claude-plugin (35 stars, last pushed yesterday), licensed Apache-2.0. It adds 56 tokens to every session and 1,367 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-04.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
next-cache-components-optimizer
Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…
next-partial-prefetching-adoption
Turn on Partial Prefetching in a Next.js app and work through the insights it surfaces. Use when the user wants to enable or adopt Partial Prefetching, flip the partialPrefetching flag, opt routes in with export const prefetch = 'partial', audit Link prefetch={true} behavior, preserve existing prefetched UI with…
chronicle
Analyze Copilot session history for standup reports, usage tips, session search, and session reindexing. Use when the user asks for a standup, daily summary, usage tips, workflow recommendations, wants to search or find past sessions by keyword/file/PR, wants to reindex their session store, or asks about deleting…