ikkeseb

ikkeseb/skills

Plugin Claude Code

Personal marketplace for the ikkeseb-skills plugin — a small bundle of day-to-day agent skills.

ikkeseb/skills

Plugin Claude Code

A small bundle of Claude Code skills for day-to-day work.

ikkeseb/skills

Instructions file CodexOpenCode

Instructions for ikkeseb/skills, covering agents.md, instruction source, structure, conventions and maintainer sessions.

ikkeseb/skills

Instructions file

Instructions for ikkeseb/skills, a project described as: Claude Code skills.

ikkeseb/skills

Agent

Adapter that runs exactly one non-interactive Codex CLI worker (OpenAI model lane) through the orchestrate skill's codex-worker.sh helper and relays its JSON result verbatim. Give it a task prompt plus model/effort/sandbox/workspace parameters and optionally a JSON Schema for the result. Not for interactive Codex…

ikkeseb/skills

Agent

Worker pinned to opus @ high for a single delegated stage that needs harness tools (implementation, research, review) without a workflow tree. Give it a complete specification with acceptance criteria and output bounds. Not for adapter or relay work (codex-worker), and not a substitute for the seat's own design…

ikkeseb/skills

Agent

Worker pinned to opus @ medium for a single tightly specified stage (bounded implementation against explicit acceptance criteria, mechanical refactors, test updates) without a workflow tree. Give it a complete specification with acceptance criteria and output bounds. Not for open problems or verification (opus-high…

ikkeseb/skills

Agent

Worker pinned to opus @ xhigh for a single delegated stage where a miss is expensive (exhaustive inventory, verification of Codex-produced work, adversarial review) without a workflow tree. Give it a complete specification with acceptance criteria and output bounds. Not for routine implementation (opus-high) or…

ikkeseb/skills

Skill Claude CodeCodex

Convert or repair a repo so AGENTS.md owns agent guidance and CLAUDE.md is a one-line adapter. Use on existing instructions, including partial conversions. Not for authoring guidance from scratch or auditing context architecture (context-audit).

ikkeseb/skills

Skill Claude CodeCodex

Audit how a project's instructions and context load across harnesses — CLAUDE.md, AGENTS.md, skills, nested files, hooks — and how its document skeleton routes each truth to one owner. Analysis only, ending in an honest worth-it verdict. Not for mechanical AGENTS.md conversion (agents-md-convert).

ikkeseb/skills

Skill Claude CodeCodex

Create a native .drawio diagram (flowcharts, architectures, ER, sequence, network topologies) that opens directly in app.diagrams.net. Not for hand-drawn style (excalidraw) or numerical charts.

ikkeseb/skills

Skill Claude CodeCodex

Create an editable .excalidraw diagram for flows, relationships, and system structure, with structural validation and visual review. Not for draw.io files (drawio) or numerical charts.

ikkeseb/skills

Skill Claude CodeCodex

Posture for exploring several independent approaches in parallel and converging on the best result, once resources are authorized. Single-task or sustained for the session.

ikkeseb/skills

Skill Claude CodeCodex

Write a paste-ready handoff for continuing the current task in a fresh session, saved temporarily and returned in the reply. Use when asked to hand off or produce continuation context. Not for reading or resuming from a pasted handoff.

ikkeseb/skills

Skill Claude CodeCodex

Mine this machine's own agent-session history (Claude Code transcripts, Codex logs) for the mistakes agents actually made, counted per model and harness, and propose per-line instruction fixes that each cite the run that earned them. A deliberate occasional audit. Not for reviewing code, a single session, or live…

ikkeseb/skills

Skill Claude CodeCodex

Adversarial-review posture for high-stakes or irreversible work: test whether the result answers the real goal and survives attempts to break it. Single-task or sustained for the session.

ikkeseb/skills

Skill Claude CodeCodex

Delegation posture: the main loop keeps design, specification, review, and integration and routes bounded, reviewable execution and reconnaissance through Claude and Codex workers — cheap models fan out by default, workhorses build and review, the seat decides. Single-task or sustained for the session. Not for one…

ikkeseb/skills

Skill Claude CodeCodex

Art-direct a single-file HTML page with editorial ambition, designing a fresh visual concept from the content each time: launches, showcases, announcements, flagship reports. Not for plain reading documents (pretty-html), decks (pretty-slides), or web-app frontends.

ikkeseb/skills

Skill Claude CodeCodex

Produce a polished, self-contained HTML reading document with light and dark themes: briefs, reports, meeting notes, prep docs, or restyling an existing HTML file. Not for showcase pages (prettier-html), decks (pretty-slides), or web-app frontends.

ikkeseb/skills

Skill Claude CodeCodex

Create designed PDFs with WeasyPrint (HTML+CSS to PDF): reports, letters, invoices, CVs, or re-typesetting a docx. Not for reading, merging, splitting, or form-filling existing PDFs, fillable forms, or quick throwaway output.

ikkeseb/skills

Skill Claude CodeCodex

Build a presentation as one self-contained HTML file: bundled slide engine, inlined fonts and images, re-skinnable through a token block. For talks and decks that stand in for PowerPoint. Not for documents (pretty-html), PDFs (pretty-pdf), or web-app frontends.

ikkeseb/skills

Skill Claude CodeCodex

Send existing work (a design, diff, diagnosis, or claim) to an OpenAI model through the Codex CLI for one independent read-only review, synthesized back. The work leaves the machine. Not for delegating execution (orchestrate).

ikkeseb/skills

Skill Claude CodeCodex

Audit factual claims in prose against traceable sources and classify each as verified, unverified, or contradicted. Checks written assertions, not software behavior.