with one click
Harvest agent-tooling ideas from prominent developers.
npx skills add https://github.com/bendrucker/claude --skill agent-ideasCopy and paste this command into Claude Code to install the skill
Harvest agent-tooling ideas from prominent developers.
npx skills add https://github.com/bendrucker/claude --skill agent-ideasCopy and paste this command into Claude Code to install the skill
| name | agent-ideas |
| description | Harvest agent-tooling ideas from prominent developers. |
| allowed-tools | ["Bash(bun ${CLAUDE_SKILL_DIR}/scripts/fetch.ts:*)","Read","Grep","Glob","WebFetch","Task","Write","Skill(things:inbox)","mcp__claude_ai_Zapier__things_create_to-do","mcp__claude-in-chrome__tabs_context_mcp","mcp__claude-in-chrome__tabs_create_mcp","mcp__claude-in-chrome__navigate","mcp__claude-in-chrome__get_page_text"] |
Harvest agent-tooling ideas from a curated list of thinkers and map them to concrete artifacts in this repo (a skill, hook, setting, or rule). The harvest runs in two phases: a feed pass that fetches, mines, and delivers a digest, and a local pass that fills the X-only gap and files the keepers. As the weekly routine these run headless-then-local, bridged by a teleport doorway item. Run on-demand at your own terminal, they collapse into one pass with no doorway (see Deliver the Doorway Item).
sources.ts is the prunable core: a version-controlled list of thinkers, each with a feed URL, or marked x-only when no usable feed exists. Edit it directly to add, drop, or retune sources. Keep it small and high-signal; a source earns its place by regularly producing ideas that map to artifacts in this repo.
Pull the last 8 days of posts across all feed sources:
bun ${CLAUDE_SKILL_DIR}/scripts/fetch.ts > tmp/agent-ideas-posts.json
Each result is {source, sourceType, feedUrl, posts:[{title,url,date,excerpt}], error?}. Sources with error set could not be fetched; note them but don't block. x-only sources are absent here by design (the local phase handles them after teleport).
Narrow to specific sources or widen the window with flags:
bun ${CLAUDE_SKILL_DIR}/scripts/fetch.ts --days 14 --source simon --source mario
Turn posts into repo-mappable idea cards. The bar is high: a card must name a real surface in this repo it would refine, or make a clear case for a net-new artifact. See references/mining.md for the card schema, the proven heuristics (per-source cap by depth, auto-demote 404 sources, dedupe against existing repo surfaces), and the fan-out flow.
Fan out one mining Task agent per source (batch thin sources together). Give each the source's posts, the schema, the heuristics, and a read-only view of the repo so it can fill repoOverlap and dedupe. Collect the cards, verify each sourceUrl resolves (demote 404s), then dedupe across agents.
Write the digest to tmp/agent-ideas-digest-<YYYY-Www>.md:
Keep the digest in this session's context as well. The local phase reads it from here.
Skip this section on a local on-demand run. The doorway exists only to bridge the headless weekly routine to a local terminal; claude --teleport ${CLAUDE_SESSION_ID} minted from a session you are already in points at itself. When the browser and Things are reachable (you invoked the skill yourself), go straight from Synthesize to Fill the X-Only Gap. The doorway is the remote routine's handoff, nothing more.
For the weekly routine, create exactly one Things inbox item. It is the routine's only notification, so make its notes self-sufficient:
Title: [agent-ideas] Weekly digest: N ideas (YYYY-Www)
Notes: the top-line summary, then the teleport command on its own line so it's tappable:
claude --teleport ${CLAUDE_SESSION_ID}
Pick the delivery path by environment:
things:inbox skill. Leave it untagged, since the local phase adds claude-code per keeper later.mcp__claude_ai_Zapier__things_create_to-do. It reaches Things through Anthropic's proxy. Tags aren't available remotely, which is fine: the doorway item is intentionally untagged.Do not tag the doorway item claude-code. Tagging happens per keeper in the local phase, so the whole digest doesn't land in the improve-claude-code backlog as one blob.
The local phase fills the X-only gap and files the keepers. You reach it two ways. The weekly routine gets here when you tap the doorway item: claude --teleport pulls the cloud session into a local terminal where the browser and Things are reachable. A local on-demand run flows straight here from Synthesize, with no doorway in between.
The digest is already in this session's context. After a teleport into a fresh terminal it may not be, so read the latest tmp/agent-ideas-digest-*.md to recover the feed-derived cards.
x-only sources were skipped in the remote fetch. Read their recent timelines in the browser and mine them into cards with the same bar as the feed path. See references/browser.md for the tool-loading sequence, the per-author flow, and the failure modes (extension not connected, logged-out wall). If the browser is unavailable, note which authors went unread and proceed with the feed cards.
For each keeper card, create one Things todo via the things:inbox skill, tagged claude-code so improve-claude-code picks it up:
[agent-idea] <card title>artifactType. This is the shape improve-claude-code consumes (short title, full notes).--tag claude-code and --session-id ${CLAUDE_SESSION_ID} so the capture carries the right tag and session attribution.Capture one todo per keeper, not one blob. Each becomes an independently triageable backlog item. Don't re-capture the feed cards as new ideas; the digest already framed them. The local phase adds the X-only cards and tags everything.
Once the keepers are captured, the work belongs to improve-claude-code: it triages the claude-code backlog into batched PRs. Tell the user how many todos landed and that they can run improve-claude-code to process them.
claude --teleport <session>, which pulls the cloud session into a local terminal with local tools. --resume does not do this. If the browser or Things isn't reachable after teleport, you may not actually be in a local session. Confirm before assuming the fallback is broken.claude-code tag is the contract with improve-claude-code. Without it, the todo won't surface there.Follow up on review feedback you received on a PR/MR: check resolution state, find silent resolves, draft replies. With --auto, autonomously triage AI-reviewer (bot) threads and loop until the reviewer is satisfied. Use when checking what review comments need responses, investigating how threads were resolved, drafting replies, or clearing a bot review hands-off.
Live tmux dashboard for reviewing inbound pull requests across GitHub and GitLab. Use when reviewing multiple PRs, checking review queue, batch reviews, or managing a review dashboard.
Monitor GitHub Actions runs and extract failure diagnostics. Use when watching PR CI, branch builds, or specific workflow runs.
Investigate GitLab CI pipeline failures and extract diagnostics. Use when watching MR CI, branch builds, or specific pipelines.
Working with GitLab merge requests via glab. Use when creating, updating, reviewing, or merging MRs.
Interactive daily review workflow across Calendar, Things, GitHub, and Linear. Use when the user asks for a daily review, morning review, evening review, or weekly review.