All authors
Claude Skills by sidhanthapoddar99
github.com/sidhanthapoddar9922 skills0 installs3 views
- SubtasksThe `issues` skill is the operating manual that lets an AI agent treat this tracker as its own queue. It must cover *traversal, reading, writing, the review handoff,* and *agent-log discipline* — and ship with helper Node scripts so the agent can filter / scan without dumping every file into context. > **History:** the spec for this skill originally lived in `2026-04-10-issues-layout/subtasks/06_documentation-and-skills.md` alongside the issues-layout user/dev docs. The docs portion was absor...Votes: 0GitHub stars: 3
- SubtasksCross-cutting authoring skill that covers markdown content itself, independent of content type. Supersedes / expands the existing `docs-guide`.Votes: 0GitHub stars: 3
- SubtasksNew skill covering the `docs` content type's file / folder / settings shape (not the markdown content itself — that's `docs:writing`).Votes: 0GitHub stars: 3
- SubtasksNew skill covering the `blog` content type's file / folder / frontmatter shape. Blog-specific (not shared markdown basics — those are `docs:writing`).Votes: 0GitHub stars: 3
- SubtasksSite-level configuration skill. Supersedes / expands the existing `docs-settings`. Covers everything *above* the per-content-type layer — the site chrome, routing, theming, aliases.Votes: 0GitHub stars: 3
- SubtasksA Claude skill that walks the user through deploying the framework in the common hosting topologies, plus skills for spinning up the canonical content types.Votes: 0GitHub stars: 3
- SubtasksStrip every reference to the `done:` subtask field from the `documentation-guide` skill — **completely, as if it never existed.** Do **not** relabel it "legacy" and do **not** leave a "this used to exist" note: any mention (even a deprecation note) still costs an agent context on every read, which is the whole reason to remove it. `state:` is the sole canonical field.Votes: 0GitHub stars: 3
- Subtasks`documentation-guide` triggers on "I'm doing docs/tracker work." It will **not** fire when an agent is heads-down in a loop thinking "I should record this iteration." That trigger gap means the structured agent-log workspace (subtask `040`) won't get used precisely when it matters most — during dense autonomous execution. Close the gap with a second, small skill that has a **different trigger surface**. **Shape: a thin *operating* skill, not a pure pointer.** A redirect-only skill costs an ex...Votes: 0GitHub stars: 3
- SubtasksCreate a new in-repo skill that teaches an agent to **build artifacts well** — distinctive, accessible, dual-theme HTML pages published as files in this repo. The skill converges four captured design skills (`artifact-design`, `dataviz`, `design-sync`, `frontend-design`, currently at repo-root `tmp_skills/`) into one self-contained skill, **written in** — adapted and rewritten for our publish mechanism, never pasted verbatim (the sources are Anthropic-authored and this repo is public). On top ofVotes: 0GitHub stars: 3
- Agent Ks ArtifactsBuild a self-contained HTML artifact in an agent-knowledge-system project — a report page, a dashboard, a chart or data visualization, a design-system or brand-guideline page, or a variation set of design options to choose between. It ships as one `.html` page in a docs section, or in an issue's `notes/` or `brainstorm/`. Trigger whenever the user asks to build, design, generate, visualize, mock up or prototype a page, dashboard, chart, UI or set of options, even when they never say artifact....Votes: 0GitHub stars: 3
- Agent Ks BlogUse this skill for blog posts in an agent-knowledge-system project — writing a new post, its YYYY-MM-DD- filename and date, frontmatter (title, description, date, author, tags, image, draft), a post's images and its assets folder, hiding or unpublishing a post with draft, and the generated blog index. Trigger it for any file under the project's blog/ folder, and whenever the user says post, article, announcement, write-up, blog or draft, or asks to publish or hide one. Markdown mechanics come...Votes: 0GitHub stars: 3
- Agent Ks CliUse the standalone agent-ks Rust CLI to navigate, search, validate and edit an agent-knowledge-system project. Covers toolkit installation, project context, issue briefs and file trees, issue/doc/blog queries, plans and log scaffolders, link-aware moves, image optimization, theme tokens, git metadata and starting the viewer. Load before running agent-ks, when a command fails, or when adding a command. The CLI has an independent GitHub release and requires no JavaScript runtime for content ope...Votes: 0GitHub stars: 3
- Agent Ks ConfigSet up and configure an agent-knowledge-system project. Use it for a new project from the starter template, a new top-level section, and every file under config/ (site.yaml, navbar.yaml, footer.yaml), the framework .env and its port, path aliases and routes, themes and theme.yaml — colours, fonts, dark mode, the logo and favicon — layout styles and custom layouts, custom pages (home, about, countdown) and their YAML, and format migrations when the site refuses to start with an engine-version ...Votes: 0GitHub stars: 3
- Agent Ks DocsUse this skill for pages inside a docs section of an agent-knowledge-system project — markdown pages, NN_ prefixes, a folder's settings.json, frontmatter, relative links, callouts, mermaid, graphviz, excalidraw and draw.io diagram pages, artifact pages, the sidebar and the outline, screenshots and image optimization, and moving or renaming a page or a folder. Trigger it for any file under a docs section of data/, for "add a doc", "write documentation", "fix this link", "shrink these screensho...Votes: 0GitHub stars: 3
- Agent Ks Index CheckCheck whether an index in an agent-knowledge-system tracker has gone stale — whether a plan, an agent log, a subtask group, issue.md or a whole issue folder still says what is true on disk. Use it whenever someone asks if a plan is out of date, whether an issue folder can still be trusted, what a folder claims against what is there, or asks for an index sweep before a wrap-up or at the start of a session that inherited someone else's work. Reports only, never edits, so it is safe to run on an...Votes: 0GitHub stars: 3
- Agent Ks Issue LogsUse for any agent log in an agent-knowledge-system issue tracker — the working folder of one run, under an issue's agent-log/. Covers when a run earns a log and when it does not, who decides, the six kinds (lp loop, rf refactor, au audit, re research, it iteration, wf workflow), the file shape of each, what a log never holds, the run statuses, and the commands that scaffold one. Trigger on agent log, loop, overnight or long-horizon run, audit, several reviewers on one target, research over ma...Votes: 0GitHub stars: 3
- Agent Ks IssuesUse for any work in an agent-knowledge-system issue tracker (data/todo/, or any folder-per-issue tracker) — issues, subtasks, plans and stages, comments, brainstorms, notes, agent memory, glossaries, the tracker vocabulary, the review queue and the issue dump. Also use for a discussion on a tracked issue, and whenever you must remember something for the next session on an issue. Trigger on issue, ticket, subtask, tracker, backlog, priority, component, label, status, plan, stage, or any file u...Votes: 0GitHub stars: 3
- Agent Ks QnaScope a subtask or a plan stage by question and answer before an agent runs it alone. Use it whenever the user starts defining a subtask or a plan, says "let me explain", "scope this", "before you start" or "this will run unattended", dictates a long story about a job, or when a subtask lacks a why, guardrails, a done-when or decisions with reasons. It writes the answers into the subtask's own sections, reason included, so an autonomous loop never stops to ask. For the shape of a subtask file...Votes: 0GitHub stars: 3
- Agent Ks Quick Idea NotePark a half-formed idea in an agent-knowledge-system tracker's issue dump — one subtask, no issue folder. Use it whenever the user wants a thought captured rather than worked on — "park this", "note this down for later", "add it to the backlog", "before I forget", "we should probably do X someday" — or when an idea has no home and no component named yet. For structural tracker work (creating an issue, plans, agent logs, editing an existing subtask) use agent-ks-issues instead. Invoke it with ...Votes: 0GitHub stars: 3
- 01 Video Artifact Engine**A skill called `agentks-video` teaches an agent to make a good explainer, not only a valid file.** It teaches a fixed order of work: the message, then the narration, then one recipe per slide, then placement, then actions tied to spoken words, then the check and the timeline, and last a look at every slide. It carries a recipe table that maps each kind of message to a slide template, item kinds and presets, and hard pacing numbers. It teaches when a video is one file and when it is a folder...Votes: 0GitHub stars: 3
- 01 Video Build- [ ] T6 and T7a first. - [ ] Then T7b and T8. Not started. T1 and T2 are built, so T6 and T7a now wait for T3 and the migration's server ([050/10](../../../2026-09-29-rust-core-engine-migration/subtasks/050_server/10_http-and-routes.md)). - [050/10 HTTP and routes](../../../2026-09-29-rust-core-engine-migration/subtasks/050_server/10_http-and-routes.md) — the server T6 and T7a need. - [080/50 Islands](../../../2026-09-29-rust-core-engine-migration/subtasks/080_ui-and-client/50_islands.md) — ...Votes: 0GitHub stars: 3
- Subtasks- [ ] **`plugins/agentks/skills/agentks-video/SKILL.md`**, about 2,000 tokens: triage, the order of work, the never-table, the trigger description. - [ ] **References:** `format.md` (both forms and who owns what, every key, the action grammar, anchors, errors), `recipes.md` (message to slide recipe, pacing numbers, motion rules), `writing-for-the-ear.md`. - [ ] **Three examples**, each passing `agentks check video`: the product tour as a folder, a concept explainer as a folder with one compon...Votes: 0GitHub stars: 3