Agent Skills

Seven Agent Skills that let a coding agent answer operational questions about Kafka through fh. Each skill collects with a bundled script that emits one fixed JSON document, reasons with stated rules, and reports in a fixed shape, so two runs on the same cluster produce the same kind of answer.

Install

The skills ship inside the fh binary. Install fh first, configure it for the deployment the agent should look at, then:

fh skills install                    # ~/.agents/skills: Codex CLI, GitHub Copilot
fh skills install --agent claude     # ~/.claude/skills: Claude Code
fh skills install --agent project    # ./.agents/skills, committed with a repository
fh skills install --dir <path>       # any other location
fh skills list                       # what this build bundles

--force replaces skills already installed. Re-run after upgrading fh; the bundled skills match the commands of the build that installed them.

Agent	Loads from	Notes
Claude Code	~/.claude/skills, .claude/skills	Skills appear as /kpow-…; kpow-replay-topic runs only when named
Codex CLI	~/.agents/skills, .agents/skills	/skills lists them; $kpow-… invokes one
GitHub Copilot (VS Code, CLI, cloud agent)	~/.agents/skills, .github/skills, .agents/skills	Picked up automatically
Other agents	the installed directory	Add one line to the agent's instructions file (for example AGENTS.md) pointing at it

The skills run fh -o json and jq; both must be on PATH.

Skills

Skill	Ask it
kpow-topic-overview	"Five biggest topics?" "Tell me about topic X." "Which topics are idle?"
kpow-consumer-health	"Is group X healthy?" "Is X keeping up?"
kpow-lag-triage	"Which consumers are behind or imbalanced?"
kpow-why-no-data	"Why isn't service S getting data from topic T?"
kpow-connect-status	"Are my connectors healthy?" "Why did X fail?"
kpow-lineage	"Where does topic X come from, who uses it?" "Map my replication."
kpow-replay-topic	"Replay topic T for group G from the beginning."

kpow-replay-topic changes committed offsets. It runs only when named explicitly, plans before it acts, and shows the plan first.

What a run looks like

The agent runs the skill's collect.sh, which makes a handful of fh calls and returns one JSON document. It applies the skill's rules to that document and writes the report in the skill's fixed skeleton, ending with a single verdict (healthy, slow, stalled, not running, hot partition) and the evidence for it. Nothing is changed unless you asked for a change.