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 agentmods add skills/dente22/spokes/skillnpx skills add Dente22/Spokes --skill skillgit clone --depth 1 https://github.com/Dente22/SpokesWhat 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 | $0.00072 | $0.00685 |
| Opus 5 | $0.00036 | $0.00342 |
| Sonnet 5 | $0.00014 | $0.00137 |
| Haiku 4.5 | $0.00007 | $0.00068 |
Grade A, and why
spokes 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 yesterday.
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 — 66 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Spokes — reuse your repos first
Spokes is a local MCP memory of the user's GitHub repositories. Treat it as a cheat sheet: if a working piece already exists in their repos, take/adapt it into the active project instead of inventing from scratch.
Always-on workflow
For every non-trivial implementation task:
- Call
start_taskwith a short description of the user request. Keep the returnedtask_id. - Call
search_reposwith the task keywords andtask_id. - If matches exist:
get_repo_cardon the best candidateslist_tree/search_in_repo/get_fileto pull only the needed pieces- Adapt into the active project
- Call
log_stepfor every action so Reuse flow shows the complete chain. - For reuse, always include:
repo_full_name+ sourcepathsource_start_line/source_end_linewhen knowntarget_pathin the active projecttarget_start_line/target_end_linewhen knowndetail: what changed and why
- If no useful match: call
log_stepwith actionskip_no_match, then implement normally. Do not force bad reuse.
Rules
- Prefer Spokes-backed reuse over guessing APIs/patterns the user already owns.
- Keep context small: fetch files/ranges, not whole repositories.
- Pass
task_idinto tools whenever possible so the web journal stays accurate. - Do not combine several actions into one
log_step. Ten actions must be ten steps in Reuse flow. - Never print the user's GitHub access token.
- If tools say GitHub is not connected, tell the user to open
http://127.0.0.1:3847, click Connect GitHub, and approve Spokes in their browser.
Tool map
| Need | Tool |
|---|---|
| Open audit task | start_task |
| Find candidate repos | search_repos |
| Inspect brief memory | get_repo_card / list_repos |
| Browse files | list_tree |
| Read code | get_file |
| Search inside one repo | search_in_repo |
| Record reuse decision | log_step |
| Read journal | get_task_journal |
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.
- yesterday First seen · 66 lines · 72 tokens per session scan A 32ef92122b34
spokes is a skill published in the GitHub repository Dente22/Spokes (1 stars, last pushed 13d ago), licensed MIT. It adds 72 tokens to every session and 685 once invoked, about $0.0004 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 2026-08-31.
Other skills, from other repositories
writing-guidelines
This skill should be used when the user asks to "review writing", "check documentation style", "audit interface copy", or "apply writing guidelines".
humanize
This skill should be used when writing or revising text, including responses, documentation, comments, emails, posts, reports, and messages, especially when the user asks to "humanize", "write naturally", "remove AI tone", or "avoid AI writing".
create-pr
This skill should be used when user asks to "create a PR", "make a pull request", "open PR for this branch", "submit changes as PR", "push and create PR", or explicitly invokes "create-pr".
dokploy-deploy
This skill should be used when user asks to "deploy with Dokploy", "use Dokploy Cloud", "manage self-hosted Dokploy", "deploy Docker Compose on Dokploy", "manage Dokploy databases", "configure Dokploy domains", or "look up Dokploy CLI commands".
gcloud-usage
This skill should be used when user asks about "GCloud logs", "Cloud Logging queries", "Google Cloud metrics", "GCP observability", "trace analysis", or "debugging production issues on GCP".
codex-advisor
This skill should be used when the user asks for a "GPT second opinion", wants a "cross-model review", needs to "check a plan before committing", wants another view after repeated failures, or explicitly invokes "codex-advisor".