CometWeb Agent Skills

Agent Skills with executable package and contract tests for research, product decisions, QA and release gates, installable in Cursor, Claude Code, Codex and other hosts that read SKILL.md packages.

Experimental distribution: task qualification is unestablished for the published packages. Each installed folder and ZIP carries its declared stage and experimental status; see skill maturity for the four stages and the boundary between package checks and task evidence.

This repository contains 34 reusable skill packages. Each one is a SKILL.md entry point plus the references, scripts and eval cases it needs, and every package is covered by the test suite. Skills supply instructions and output contracts. They do not supply a model, connector accounts or permission to act: browser access, connectors and external side effects stay with your host and your authorization.

• Quickstart
• Pick a starting point
• Skill catalog
• How skills hand off: CW-AIP
• Write and evaluate a new skill
• Contributing

What the tests establish

Evidence	What passing establishes	What remains unmeasured
Package gates and CI badges above	The identified source meets package, security and contract checks	Correctness of every model answer
Deterministic behavior suites	Helpers enforce their tested invariants	Live task success and comparative lift
Host adapters and compatibility matrix	Declared formats and capabilities; Claude/Codex execution adapters exist	A successful session on every host/model
Real-host evals	Only the recorded runs, exact identities and reviewed outcomes	General effectiveness without independent, repeated A/B

Live task success remains NOT_ASSESSED for the catalog as a whole. Real-host evaluation separates structural grades, blind review, cost and missing execution. A passing CI run grants no release authority.

Quickstart

Clone once, then run the installer for your host. Installers symlink each package from skills/ into the host's skills directory, so keep the clone where it is; git pull updates every installed skill in place.

git clone https://github.com/CometWeb-io/agent-skills.git
cd agent-skills
./scripts/install-claude.sh --skill evidence-researcher --skill product-operator
# Other hosts: install-cursor.sh, install-codex.sh, install-qwen.sh,
# install-qoder.sh, install-lingma.sh. Omit --skill to install every package.
# ./scripts/install-all.sh explicitly installs into all six hosts.

Installing every package ends with a line such as OK: 34 Claude Code skills installed in /home/you/.claude/skills. Installers support repeatable --skill ID and --ref <full-commit-SHA> for a clean pinned checkout. Keep stable installations on a detached reviewed commit; see pinning and rollback. Installers are safe to rerun, stop before changing anything when a target path is already taken, and accept --dry-run and --uninstall. INSTALL.md lists each host's target directory and override variable, conflict backups, upgrades, and the plugin marketplaces for Claude Code, ChatGPT and Codex.

Claude Code can install the same skills as a plugin instead, without a clone:

claude plugin marketplace add CometWeb-io/agent-skills
claude plugin install cometweb-agent-skills@cometweb-agent-skills

Then ask your assistant:

Use product-operator to compare this repo with the roadmap.
Give me the three most useful next actions and how to verify each one.

Pick a starting point

Use the owner and handoff map to choose a task owner, follow its output contract, and install a small core plus the needed domain skills.

You need to…	Start with	You get
Check a claim	evidence-researcher	Sources, contradictions and an Evidence Pack
Decide what to do next	product-operator	A bounded now / next / stop list
Inspect an app	web-app-auditor	Findings tied to observed behavior
Review a release	release-readiness	A verdict for a specific candidate
Combine specialists	skill-orchestrator	Ordered steps and structured handoffs

For isolated specialist runs, use skill-orchestrator-multiagent.

Skill catalog

34 skills. Versions come from each package's VERSION; the compatibility matrix lists host support.

Foundation and orchestration

Skill	Use it for	Version
ai-council	Evidence-governed decisions, risk gates, forecasts, and GO / TEST / DEFER verdicts. (experimental; DEVELOPMENT; runs only when named)	5.3.2
cometweb-context	Fresh, provenance-aware context snapshots before work that depends on current project state. (experimental; DEVELOPMENT)	1.6.1
evidence-researcher	Claim decomposition, source verification, falsifiers, contradictions, and Evidence Packs. (experimental; DEVELOPMENT)	1.0.11
portfolio-operator	Cross-project focus, capacity conflicts, and pause / delegate decisions. (experimental; DEVELOPMENT)	1.2.5
skill-orchestrator	Multi-skill workflows with ordered steps and CW-AIP handoffs. (experimental; DEVELOPMENT)	1.4.3
skill-orchestrator-multiagent	Isolated subagent execution for multi-skill workflows. (experimental; DEVELOPMENT)	1.2.4
benchmark-curator	Benchmark corpora, holdouts, contamination controls, and revision hashes. (experimental; DEVELOPMENT)	1.7.5
feedback-integrator	Recurring failure patterns, improvement proposals, and regression tests. (experimental; DEVELOPMENT)	1.8.2
quality-loop-operator	Briefing, review, repair, acceptance, measurement, and quality lifecycle control. (experimental; DEVELOPMENT)	1.7.7
rubric-designer	Observable evaluation criteria, evidence floors, blocker rules, and rubric locks. (experimental; DEVELOPMENT)	1.7.7
skill-auditor	Skill routing, portability, package hygiene, and supply-chain audits. (experimental; DEVELOPMENT)	1.7.6
skill-evaluator	Fair skill experiments, behavioral lift, resource cost, and host comparisons. (experimental; DEVELOPMENT)	1.8.1

Product, research, and partnerships

Skill	Use it for	Version
ai-humanize	Natural English and Polish rewrites that preserve meaning and voice. (experimental; DEVELOPMENT)	2.6.5
competitive-intelligence	Competitor watchlists, change detection, and recurring delta digests. (experimental; DEVELOPMENT)	1.1.5
design-partner-finder	Finding, qualifying, and managing design partners and early adopters. (experimental; DEVELOPMENT)	1.2.5
product-operator	Weekly product control loops, roadmap drift, and now / next / later / stop actions. (experimental; DEVELOPMENT)	2.4.4
product-teardown	Evidence-backed product, UX, architecture, and implementation pattern analysis. (experimental; DEVELOPMENT)	1.3.2
repo-to-roadmap	Whole-project baselines, gap inventories, dependencies, and target-state roadmaps. (experimental; DEVELOPMENT)	1.1.3
brief-architect	Explicit artifact contracts, evidence policies, risks, and acceptance criteria. (experimental; DEVELOPMENT)	1.8.1
content-writer	Evidence-aware reader-facing articles, guides, reports, and documentation. (experimental; DEVELOPMENT)	1.7.7
content-reviewer	Constructive editorial QA with evidence-backed, actionable findings. (experimental; DEVELOPMENT)	1.7.6
content-roaster	Adversarial content review, proof debt, objections, and repair verification. (experimental; DEVELOPMENT)	6.1.7
science-roaster	Reviewer #2-style critique of methods, inference, validity, and reproducibility. (experimental; DEVELOPMENT)	6.1.7
repo-roaster	Adversarial repository review with invariants, reachability, and repair contracts. (experimental; DEVELOPMENT)	6.1.7
repair-operator	Minimal dependency-aware repairs and fresh verification of closed findings. (experimental; DEVELOPMENT)	1.7.7
artifact-acceptance	Final evidence-backed acceptance gates for knowledge artifacts. (experimental; DEVELOPMENT)	1.7.6
competitor-profiling	Initial evidence-backed competitor baselines and normalized handoffs to competitive intelligence. (experimental; DEVELOPMENT)	0.2.1
research-program-operator	Research-program stage gates, next-study planning, and manuscript-readiness handoffs. (experimental; DEVELOPMENT; frozen)	1.3.4

Publication, operations, QA, and release

Skill	Use it for	Version
customer-ops	Support triage, incidents, account risk, commitments, and engineering handoffs. (experimental; DEVELOPMENT)	2.2.3
ebook-publisher	Research-backed ebooks, white papers, workbooks, and publication QA. (experimental; DEVELOPMENT)	1.0.6
longform-publisher	Canonical long-form manuscripts and release-ready derived documents. (experimental; DEVELOPMENT; frozen)	1.1.6
release-readiness	Candidate-bound production gates and GO / GO_WITH_CONTROLS / NO_GO / DEFER verdicts. (experimental; DEVELOPMENT)	1.3.4
seo-geo-aeo-maxxing	Multi-pillar SEO / GEO / AEO visibility audits. (experimental; DEVELOPMENT)	1.3.6
web-app-auditor	Evidence-driven click-through QA for websites and web applications. (experimental; DEVELOPMENT)	1.5.1

What is inside a skill

skills/<id>/
  SKILL.md        front door: frontmatter description (what routes it) + workflow
  VERSION         semantic version, mirrored in registry/skills.json
  references/     detail the host reads only when the skill opens it
  scripts/        deterministic kernels, validators, run_evals.py
  evals/, tests/  cases that run under pytest
  agents/         generated host adapters (do not hand-edit)

registry/skills.json is the source of truth for descriptions, versions, ownership boundaries and routing signals. The Cursor routing rule, the compatibility matrix and the catalog above are generated from it. A passing validator checks a contract; it does not prove that an AI answer is correct.

How skills hand off: CW-AIP

Skills stay standalone, but when one feeds another the handoff is a typed JSON envelope rather than prose. That is the CometWeb Agent Interchange Protocol.

question ──▶ evidence-researcher ──▶ EvidenceEnvelope ──▶ ai-council ──▶ DecisionEnvelope
web-app-auditor ──▶ FindingEnvelope[] ──▶ release-readiness ──▶ release verdict
repo-to-roadmap ──▶ roadmap baseline ──▶ product-operator ──▶ handoff to a specialist

• Small core, typed payload. Every envelope carries identity, producer, as_of time, dependencies and a payload with its hash; each payload kind has its own JSON Schema.
• Producer truth. Only the emitting skill sets producer and timestamps; a consumer that disagrees emits a new envelope instead of editing one.
• Status is local. Verified for research is not verified for a decision, which is not verified for a release. Consumers may apply stricter gates.

protocol/cw-aip-v2/ is current; protocol/cw-interchange-v1.md stays valid for skills that have not migrated. skill-orchestrator runs multi-step flows and passes the envelopes between steps. Check an envelope with:

uv run python tooling/validate_envelope.py --final fixtures/cwaip-v2/evidence-final.json

Write and evaluate a new skill

You need uv and Python 3.12+.

uv sync --group dev
uv run python tooling/new_skill.py my-skill \
  --description "80-1024 characters: what the skill does, when to use it, and when not to" \
  --summary "One line for the README catalog."
uv run python tooling/check_all.py --fast

The scaffold is registered and passes every fast gate straight away, but its content is placeholder. CONTRIBUTING.md walks through replacing it, making the evals and routing cases real, and bumping the plugin version. Full admission and versioned packaging reject known generated author prompts even if their scaffold labels were removed. Declared SCAFFOLD also blocks versioned publication; rewritten instructions can be distributed only with explicit experimental maturity metadata. A development ZIP (package_skill.py --dev) remains unassessed. Deterministic evals prove a package keeps its contract; whether a skill makes a model behave better is a separate experiment, described in docs/CONTRACT-TRACE-AND-SKILL-EVALS.md and run with the skill-evaluator skill.

Contributing

Read CONTRIBUTING.md for the check flow and the quality bar, and SECURITY.md for reporting a vulnerability. The documentation index lists everything else, from how routing works to evaluating and releasing.

MIT license · Changelog

Experimental opt-in product profiles

The PRD profile in brief-architect, four specialist profiles and three imported Corey capability packs are source previews. Default orchestrator plans do not activate them. Use an explicit PRD request, --with-prd-handoff, --specialist-profile or --with-capability-packs for the corresponding preview. Profile and PRD plans remain PLAN_ONLY; they grant no execution permission.

JSON schemas, semantic kernels, source bindings and resumable ledgers are checked with deterministic tests. Live host compatibility, output quality and comparative lift remain unqualified. Capability packs provide FRAMEWORK guidance, never independent evidence of demand or business outcomes. Retained licenses and pinned source commits are listed in THIRD_PARTY_NOTICES.md and the registries.

Repository checks validate this source candidate; they do not authorize installation, publishing or promotion of the experimental profiles.

Candidate quality boundaries

competitor-profiling and research-program-operator are experimental runtime candidates: their deterministic output contracts are tested, while real-model behavior and task-success lift remain not assessed. Catalog presence and format support are not host/model qualification. A READY handoff is bounded readiness for its named next step, never publication or external-write authority.

Discovery metadata, activated instructions and optional reference resources have separate estimates in the context budget. Actual prompt/cache tokens and runtime costs require recorded host runs.