writing-a29bcb — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited writing-a29bcb (Plugin) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Plugin marketplace for Claude Code and Codex.
Bundles 12 plugins across Atlassian, Google Workspace, Databricks, agent-system management, design workflows, research, writing, terminal control, learning, and more.
| Skill | Plugin | What it does |
|---|---|---|
jira | atlassian | Search Jira issues, create and update tickets, transition workflows, comment, manage sprints, run bulk operations |
confluence | atlassian | Search Confluence pages, read documentation, create and update pages, browse spaces |
gmail | google-workspace | Triage inbox, search and read messages, send mail, manage drafts, labels, and filters via the gws CLI |
calendar | google-workspace | View agenda, manage events, check availability, manage calendars via the gws CLI |
research | research | Research complex topics and produce sourced reports |
writing | writing | Draft, review, and finish long form prose |
pyramid | writing | Structure analytical documents with the Pyramid Principle |
tech-doc | writing | Draft, review, and finish technical documentation |
presentations | writing | Design slide-deck content (audience, message, storyboard, per-slide briefs); hands off to crafting-presentations for HTML |
claude-codex-bridge | runtime-bridge | Align Claude Code and Codex project files |
improving-instructions | agent-system-management | Audit and improve agent instruction files |
capturing-session-learnings | agent-system-management | Capture session learnings into the right instruction file |
creating-skills | agent-system-management | Scaffold, iterate, pressure-test, and tune skills across the full lifecycle |
brainstorming | workbench | Sequential Q&A to clarify design intent |
writing-spec | workbench | Synthesize a design discussion into a spec doc |
writing-plans | workbench | Turn approved specs into concrete implementation plans |
visualizing-options | workbench | Browser-based visual companion for layout choices |
using-workbench | workbench | Load Workbench skill rules and routing |
terse-mode | workbench | Compact token-saving replies until disabled |
autopilot | workbench | Ship a feature from brainstorm to PR using a project profile |
copilot | workbench | Human-in-the-loop autopilot: you drive brainstorm and spec, the agent ships the rest |
verification-before-completion | workbench | Require fresh verification evidence before completion claims |
test-driven-development | workbench | Enforce test-first RED-GREEN-REFACTOR implementation discipline |
dispatching-parallel-agents | workbench | Split independent tasks across concurrent agents |
subagent-driven-development | workbench | Execute implementation plans with fresh agents and review gates |
systematic-debugging | workbench | Root-cause investigation before proposing bug fixes |
crafting-html | workbench | Reference gallery of 21 HTML artifact patterns |
crafting-design-systems | workbench | Design systems (CSS variables, components, images) that theme HTML producers |
crafting-presentations | workbench | Multi-slide HTML decks with deck navigation, slide-type catalog, and a presenter view |
exporting-decks-to-pptx | workbench | Export a deck-stage HTML deck to a native, editable PowerPoint (.pptx) via python-pptx |
perfecting-presentations | workbench | Strict deck review loop: hard gates plus adversarial judge panels until rounds come back dry |
tmux | terminal | Control interactive terminal programs through isolated tmux sessions |
frontend-design | frontend-design | Distinctive, production-grade frontend interfaces that avoid generic AI aesthetics |
emil-design-eng | frontend-design | Emil Kowalski's design engineering philosophy: animation timing, component polish, UI craft |
playground | playground | Interactive single-file HTML playgrounds with controls, live preview, and copy-out prompt |
databricks-core | databricks | Databricks CLI, authentication, profile management, and data exploration |
databricks-docs | databricks | Live docs.databricks.com lookups for product-surface questions |
quizzing-the-session | learning | Get taught and quizzed on the current session's work until you demonstrably understand the problem, solution, and impact |
quizzing-a-topic | learning | Get taught and quizzed on any topic or theme you name until you demonstrably understand it |
Skills are invoked from the host agent (Claude Code or Codex) using the fully qualified form /<plugin>:<skill>, for example /atlassian:jira or /workbench:autopilot.
/plugin marketplace add pgoell/pgoell-claude-tools
/plugin install atlassian@pgoell-claude-tools
/plugin install google-workspace@pgoell-claude-tools
/plugin install research@pgoell-claude-tools
/plugin install writing@pgoell-claude-tools
/plugin install runtime-bridge@pgoell-claude-tools
/plugin install agent-system-management@pgoell-claude-tools
/plugin install workbench@pgoell-claude-tools
/plugin install terminal@pgoell-claude-tools
/plugin install frontend-design@pgoell-claude-tools
/plugin install playground@pgoell-claude-tools
/plugin install databricks@pgoell-claude-tools
/plugin install learning@pgoell-claude-toolsAdd the marketplace from your shell, then install plugins from inside Codex:
codex plugin marketplace add pgoell/pgoell-claude-tools
codex
/pluginsIn the /plugins picker, install any combination of atlassian, google-workspace, research, writing, runtime-bridge, agent-system-management, workbench, terminal, frontend-design, playground, databricks, and learning.
To pick up updates: codex plugin marketplace upgrade pgoell-claude-tools and re-install the affected plugins.
Jira and Confluence skills for Atlassian Cloud.
Skills:
/atlassian:jira: Search issues, update tickets, transition status, add comments, and manage sprints./atlassian:confluence: Search pages, read documentation, update pages, and browse spaces.Setup:
The plugin supports two authentication paths.
Atlassian CLI (recommended).
brew install atlassian/tap/acli
acli auth loginAPI token fallback.
Generate a token at https://id.atlassian.com/manage/api-tokens, then export:
export ATLASSIAN_DOMAIN="your-domain" # e.g. mycompany (for mycompany.atlassian.net)
export ATLASSIAN_EMAIL="[email protected]"
export ATLASSIAN_API_TOKEN="your-token"Gmail and Calendar skills for Google Workspace, powered by the gws CLI.
Skills:
/google-workspace:gmail: Search, read, send, and manage Gmail messages, drafts, labels, and filters./google-workspace:calendar: View agendas, manage events, check availability, and manage calendars.Setup:
Install and authenticate the gws CLI:
npm i -g @anthropic-ai/gws
gws auth login -s gmail,calendarFull setup instructions: https://github.com/googleworkspace/cli
Research complex topics and produce sourced reports.
Skills:
/research:research: Plan focused investigations, gather sources, synthesize findings, review conclusions, and write reports. Reports default to HTML at reports/<topic-slug>-<YYYY-MM-DD>/report.html; override per invocation with "give me a markdown research report".Writing skills for prose, analytical structure, and technical documentation.
Skills:
/writing:writing: Draft, review, and finish long form prose./writing:pyramid: Structure memos, recommendations, briefings, and decision documents with the Pyramid Principle./writing:tech-doc: Draft, review, and finish tutorials, how-to guides, references, and explanations./writing:presentations: Design slide-deck content end to end (intake, message architecture, storyboard, per-slide drafts, critique panel). Produces a markdown deck.md you can hand off to /workbench:crafting-presentations for HTML rendering. Also runs in audit mode against an existing deck.md.Aligns Claude Code and Codex project configuration.
Skills:
/runtime-bridge:claude-codex-bridge: Align project files, settings, hooks, agents, and plugin availability across the two runtimes.Manage the host agent's instruction layer and skill layer: audit AGENTS.md / CLAUDE.md, capture session learnings, and scaffold or iterate on Claude Code / Codex skills.
Skills:
/agent-system-management:improving-instructions: Audit and improve AGENTS.md and CLAUDE.md files./agent-system-management:capturing-session-learnings: Capture session learnings into the right instruction file./agent-system-management:creating-skills: Scaffold a new skill, iterate on an existing one with eval loops, pressure-test discipline skills, optimize triggering, or extract a skill from a conversation.Workbench skills for design dialogue, skill routing, and profile-driven feature shipping.
Skills:
/workbench:brainstorming: Sequential question-and-answer loop to clarify design intent. Hands off to writing-spec when the design is ready to be written down./workbench:writing-spec: Synthesize a design discussion into a spec doc, run a fresh-eyes self-review subagent, gate on user approval, then hand off to workbench:writing-plans./workbench:writing-plans: Turn approved specs into concrete implementation plans./workbench:visualizing-options: Browser-based visual companion for mockups, layout comparisons, wireframes, and architecture diagrams./workbench:using-workbench: Load Workbench skill rules and routing./workbench:terse-mode: Compact token-saving replies until disabled./workbench:autopilot: Ship a feature from brainstorm to PR using a project profile. Profile schema documented in plugins/workbench/skills/autopilot/references/profile-schema.md./workbench:copilot: Human-in-the-loop sibling of autopilot. You drive the brainstorm and spec by hand; the agent automates plan, implementation, docs, audit, and PR. Reads the same .workbench/autopilot.md profile as autopilot./workbench:verification-before-completion: Require fresh verification evidence before completion claims./workbench:test-driven-development: Enforce test-first RED-GREEN-REFACTOR implementation discipline./workbench:dispatching-parallel-agents: Split independent tasks across concurrent agents./workbench:subagent-driven-development: Execute implementation plans with fresh agents and review gates./workbench:systematic-debugging: Enforce root-cause investigation before proposing bug fixes; bundles techniques for backward stack tracing, defense in depth, and condition-based waiting./workbench:crafting-html: Reference gallery of 21 HTML artifact patterns vendored from ThariqS/html-effectiveness. Activates for standalone HTML artifacts not covered by specs, plans, brainstorm summaries, debug reports, or research reports./workbench:crafting-design-systems: Create reusable design systems (CSS variables, components, images) at project (.workbench/design-systems/<name>/) or user (~/.claude/workbench/design-systems/<name>/) scope. HTML producers inline the active design system over their template defaults./workbench:crafting-presentations: Multi-slide HTML decks with a deck-stage engine, slide-type catalog (Title, SectionDivider, Agenda, Content, Stat, Capabilities, Comparison, Quote, Timeline, Closing), speaker notes JSON island, and a two-window presenter view that syncs over BroadcastChannel, hides the deck's editing chrome while presenting, and lets you edit notes live as markdown and save them back into the deck file. Bundles a Deloitte/Databricks Alliance example deck./workbench:exporting-decks-to-pptx: Convert a finished deck-stage HTML deck (single-file or multi-file) into a native, editable PowerPoint file via a freshly written python-pptx generator, with a dockerized LibreOffice render-verify loop and an optional per-slide adversarial verification panel./workbench:perfecting-presentations: Strict review loop for existing HTML decks. Deterministic hard gates (overflow, overlap, contrast, broken assets, typography lint) plus four fresh-context judge dimensions (narrative, clarity, visual, delivery) whose findings must cite a per-project constitution (.workbench/deck-standards.md, seeded from a bundled default) and survive adversarial verification. Terminates on convergence: two consecutive dry rounds, with a rejected-findings ledger to prevent oscillation.writing-spec, writing-plans, brainstorming, and systematic-debugging can emit either markdown or HTML; defaults are markdown for specs and plans, HTML for brainstorm summaries and debug reports. Override per invocation or via .workbench/config.md (schema in plugins/workbench/skills/autopilot/references/config-schema.md).
Terminal skills for interactive command-line programs.
Skills:
/terminal:tmux: Drive interactive CLIs, REPLs, debuggers, and experimental nested agent sessions through isolated tmux sessions.Setup:
Install tmux on Linux, macOS, or WSL. Native Windows terminals are not supported.
Distinctive, production-grade frontend interfaces with deep UI craft and animation discipline. Ports two complementary skills: Anthropic's frontend-design (Apache 2.0) for creative direction, and Emil Kowalski's emil-design-eng for design engineering and animation discipline (the upstream repo declares no license; included under the upstream author's public publishing intent). See plugins/frontend-design/NOTICE for full attribution.
Skills:
/frontend-design:frontend-design: Build web components, pages, and applications with a clear aesthetic point of view (typography, color, motion, composition)./frontend-design:emil-design-eng: Review UI craft, choose easing curves and durations, build interaction-rich components, and audit motion.Interactive single-file HTML playgrounds: control panel, live preview, and copy-out prompt. Ports Anthropic's playground plugin (Apache 2.0; see plugins/playground/NOTICE for attribution).
Skills:
/playground:playground: Build self-contained HTML playgrounds for design, data, concept maps, document critique, diff review, or code architecture.Two Databricks skills. databricks-core is a verbatim port from databricks/databricks-agent-skills at commit bf6d932, governed by the upstream Databricks License (use restricted to Databricks Services; see plugins/databricks/LICENSE-upstream and plugins/databricks/NOTICE). databricks-docs is original authorship: it points the host agent at docs.databricks.com via WebFetch using a curated URL index. The plugin also preserves references/research/ with 12 MB of Genie Code reverse-engineering notes; see plugins/databricks/README.md for details.
Skills:
/databricks:databricks-core: CLI, authentication, profile management, data exploration, and bundles./databricks:databricks-docs: Live docs.databricks.com lookups for product-surface questions (Apps, DABs, Jobs, Lakebase, Model Serving, Pipelines, Unity Catalog, SQL Warehouses, MLflow, serverless, secrets, workflows).Setup:
databricks-core operates against your existing databricks CLI profiles; configure them with databricks auth login or by editing ~/.databrickscfg. databricks-docs is read-only against docs.databricks.com and needs no auth.
Socratic teach-and-quiz skills that verify you deeply understand something. An original adaptation of Thariq Shihipar's "Learn Quiz" gist (see plugins/learning/README.md for attribution).
Skills:
/learning:quizzing-the-session: Build a problem/solution/impact checklist from the current session and recent git activity, then teach and quiz you item by item to mastery./learning:quizzing-a-topic: The same teaching engine pointed at any topic or theme you name, grounded in repo files when the topic is local code.See AGENTS.md for repository structure, plugin layout, and contribution conventions.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.