Install
$ agentstack add skill-product-on-purpose-agent-skills-toolkit-askit-capability-advisor ✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
About
askit-capability-advisor
Purpose
Tell a maintainer what their target agent supports before they build, so the plan fits the agent rather than failing at release. Two modes: advise takes the target agents (claude, codex, or both), reports which component types are plugin-distributable on each, flags the Claude-only types and the cross-agent asymmetries, and recommends a tier path; check takes a specific component or a whole plugin and reports which declared targets can actually run it, so a Claude-only component shipped under a Codex target is caught before release. The full matrix is in [references/capability-matrix.md](references/capability-matrix.md).
When to use
When choosing agent-targets, checking whether a component is portable across agents, or deciding which conformance tier to aim for.
advise mode
- Read the target agents (
claude,codex, or both). - For each, report the plugin-distributable component types and the asymmetries: subagents are Claude-only for plugin distribution (Codex ingests them only via user/project
config.toml, Standard sec 3.3); output styles are Claude-only and the statusline differs (sec 2.3); a command's Codex form is its backing skill, explicitly invocable, not a separate artifact (sec 3.2); hooks are a Codex subset of Claude's event set (sec 3.5); skills, MCP, and AGENTS.md are portable on both (sec 3.1, 3.9, 3.10). - Recommend a tier path: Bronze (skills + references + AGENTS.md + MCP), Silver (adds subagents, commands, workflows, chain contracts, plugin packaging, the prefix, native manifests), Gold (adds hooks, output styles, self-hosting CI).
check mode
- Read the component (or every component in the plugin) and the plugin's declared
agent-targets. - For each declared target, report whether the type is plugin-distributable there. Flag a Claude-only type (subagent, output style, statusline) carried under a
codextarget, since it cannot ship to Codex from a plugin. - Point each flag at the fix: set a per-component
agent-targets: [claude]override (sec 3.7), or move the component into a Claude-only plugin.
Scope
Advisory only: it reports capability and recommends a tier; it does not author components (the askit-build-* skills do) or grade conformance (askit-evaluate does). The matrix tracks the agent versions the Standard pins per component (the load-bearing Codex subagent constraint is pinned to Codex CLI v0.135 in sec 3.3); when an agent adds a capability the matrix and this skill update together, so the advice never drifts ahead of what the agents actually support.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: product-on-purpose
- Source: product-on-purpose/agent-skills-toolkit
- License: Apache-2.0
- Homepage: https://product-on-purpose.github.io/agent-skills-toolkit
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet — be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.