hep-call

SkillAI & models

Once this skill is added, your AI can prepare Agentlas Hub or Cloud agents that you name explicitly. That means you can point to a specific agent by name and have it prepared for use. The skill comes from the agentlas-os project on GitHub.

Available today. Use it from your connected AI after setup.

After adding the skill, tell your AI the name of the Agentlas Hub or Cloud agent you want prepared. It will take care of the preparation from there.

Then ask your AI: use the hep-call skill

What your AI can do with it

  • Prepare a named Agentlas Hub agent
  • Prepare a named Agentlas Cloud agent
  • Work with both Agentlas Hub and Agentlas Cloud agents
  • Target a specific agent by its exact name

What this skill tells your AI

The instructions your AI receives, as published by agentlas-ai/agentlas-os in kimi/skills/hep-call/SKILL.md and read by ahel’s review.

Update fallback: 자동 업데이트가 안 되면 hephaestus update를 한 번 실행하세요. 업데이트하지 않아도 현재 버전 명령은 그대로 동작합니다.

/hep-call

Call the exact agents named by the user. This prepares BYOM runtime bundles and receipts; your own model still performs the actual model/tool execution.

Syntax: /hep-call agent-a, agent-b {context}. The text before { is the agent list; the text inside braces is the context. If braces are omitted, the first token is the agent list and the rest is context.

If the user wants it to keep running

A Hub call is charged every time it runs — the 24-hour auto-lease was retired 2026-08-18, so paying once does not make the next call free. A job that wakes on a schedule pays on every wake-up: a five-minute watcher is 288 calls a day.

When the request is a standing or recurring one, call hephaestus.quote_agent_lease for the named agent before the first run. It answers in one shot — per-day price, total for the days considered, the workspace's current balance, and the shortfall plus a top-up link when the balance is short. Show all of it in one message and ask how many days they want; do not choose for them. Only then call hephaestus.purchase_agent_lease — a confirmed purchase needs confirm: true, the chosen days, the confirmationToken, the expectedPerDayCredits and expectedTotalCredits from the quote, and a stable idempotencyKey, or it is refused.

Never buy a lease the user did not agree to. On insufficient_credits, say how short they are and give the top-up link instead of asking them to approve a purchase that cannot go through. On leaseOffered: false the creator set no per-day price and it is genuinely not for sale — say so and quote the per-call cost instead.

How to run

Run the shell block below verbatim, replacing only the RAW value with the exact text the user typed after /hep-call. The runner is resolved by absolute path — nothing to install, nothing to add to PATH.

Guardrails: do NOT diagnose command not found/PATH, do NOT edit ~/.zshrc, and do NOT create/commit/stash/push any git branch or claim you did. This workflow changes no source files. If the runner is missing, say so and stop — never fabricate a run.

RAW="<replace with the exact text the user typed after /hep-call, e.g. agent-a, agent-b {context}>"

case "$RAW" in
  "<replace"*) echo "RAW placeholder not filled — substitute the user's request first." >&2; exit 2 ;;
esac

RUNNER=""
for candidate in \
  "$HOME/.agentlas/runtime/current/bin/hephaestus" \
  "./bin/hephaestus"
do
  if [ -n "$candidate" ] && [ -x "$candidate" ]; then RUNNER="$candidate"; break; fi
done
if [ -z "$RUNNER" ]; then
  for cache in "$HOME/.claude/plugins/cache/agentlas-core-engine/hephaestus" \
               "${CODEX_HOME:-$HOME/.codex}/plugins/cache/agentlas-core-engine/hephaestus"; do
    newest="$(ls -d "$cache"/*/bin/hephaestus 2>/dev/null | sort -V | tail -1)"
    if [ -n "$newest" ] && [ -x "$newest" ]; then RUNNER="$newest"; break; fi
  done
fi
[ -n "$RUNNER" ] || { echo "Hephaestus runtime not found. Run the installer first." >&2; exit 1; }
if [ "${HEPHAESTUS_AUTH_AUTOPOPUP:-1}" != "0" ]; then
  "$RUNNER" auth ensure >/dev/null 2>&1 || true
fi
if printf '%s' "$RAW" | grep -q '{'; then
  AGENTS="${RAW%%\{*}"
  CONTEXT="${RAW#*\{}"
  CONTEXT="${CONTEXT%\}}"
else
  AGENTS="${RAW%% *}"
  CONTEXT="${RAW#* }"
fi
AGENTS="$(printf '%s' "$AGENTS" | sed 's/[[:space:]]*$//')"
CONTEXT="$(printf '%s' "$CONTEXT" | sed 's/^[[:space:]]*//;s/[[:space:]]*$//')"
"$RUNNER" call "$AGENTS" "$CONTEXT" --runtime "${AGENTLAS_HOST_RUNTIME:-terminal}"

Answer shape

  1. Report each requested agent slug and status.
  2. If status: "prepared", use the returned output.entry_excerpt and output.grounding.directive as the agent's runtime instructions.
  3. If status: "insufficient_credits", tell the user the credits needed vs. have and point to upgrade; do NOT run the agent.
  4. If an agent fails for any other reason, continue with prepared agents and clearly list each failure with its status.
  5. NEVER substitute for a blocked, failed, or metered agent by reading its local source files and role-playing the persona yourself. A Hub agent runs only through the server's metered bundle; if it did not return prepared, report why — never fabricate a run.
  6. Include the top-level receipt_id and every prepared execution_id.

Examples

/hep-call market-researcher, report-writer {시장 리포트 초안 만들어줘}
/hep-call cloud:my-finance-agent {이 리포트 리스크 검토}

Update fallback: 자동 업데이트가 안 되면 hephaestus update를 한 번 실행하세요. 업데이트하지 않아도 현재 버전 명령은 그대로 동작합니다.

Signals

GitHub stars
1k
Forks
103
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
hep-call
Source
github.com/agentlas-ai/agentlas-os