plugins/promode/skills/managing-promode/SKILL.md
MUST be loaded when setting up, installing, updating, or auditing promode in a project. Promode delivers a main-agent orchestration brief via a SessionStart hook and never puts its methodology in the project's own CLAUDE.md. Invoke PROACTIVELY when the user mentions installing/setting up promode, the promode hook, or auditing a promode install.
npx skillsauth add mikekelly/promode managing-promodeInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
<essential_principles>
This skill sets up the promode methodology in a project. Promode never puts its methodology in CLAUDE.md (that's hook-delivered) and never clobbers it.
1. The main-agent brief is delivered by a hook, not CLAUDE.md
PROMODE_MAIN_AGENT.md (the team-lead orchestration brief) is injected into the main agent only, by a SessionStart hook gated on agent_id. It never reaches subagents. See references/main-agent-delivery.md for why.
2. CLAUDE.md is the project's own file and the agent-knowledge root
The project's CLAUDE.md is auto-loaded into every agent and serves as the root of the agent-knowledge graph. Promode does not put its orchestration brief there. Agents maintain it over time (adding links to new knowledge docs); existing content is never overwritten. A minimal CLAUDE.md may be created if none exists.
3. Subagents are self-contained Phase agents (implementer, code-reviewer, debugger, …) carry the methodology in their own definitions, so they need nothing installed into the project.
4. Agent knowledge is an interlinked markdown graph rooted at CLAUDE.md
Durable project knowledge for agents lives as a graph of linked markdown docs. CLAUDE.md is the root — it links out to key areas; docs link to each other densely. File location doesn't matter — the links carry the graph. Distinct from README.md (for humans). See references/agent-knowledge-wiki.md.
</essential_principles>
<never_do>
CLAUDE.md; NEVER overwrite or clobber existing CLAUDE.md content (the knowledge graph only appends + links; a minimal CLAUDE.md may be created if none exists)standard/PROMODE_MAIN_AGENT.md or the standard hook when copying into a project — copy them exactly.claude/settings.json — always MERGE the hook entry, preserving existing settings
</never_do>Wait for response before proceeding. </intake>
<routing> | Response | Workflow | |----------|----------| | 1, "install", "set up", "add promode" | workflows/install.md | | 2, "update", "upgrade", "refresh" | workflows/update.md | | 3, "audit", "check", "verify" | workflows/audit.md |After reading the workflow, follow it exactly. </routing>
<quick_reference> Promode install footprint in a target project:
<project>/
├── CLAUDE.md # the project's own file + agent-knowledge root (created if missing, never clobbered)
└── .claude/
├── PROMODE_MAIN_AGENT.md # main-agent brief (copied from standard/)
├── hooks/promode-main-context.sh # SessionStart hook (copied from standard/)
└── settings.json # SessionStart entry (MERGED; matchers startup|resume|clear|compact)
The hook injects PROMODE_MAIN_AGENT.md into the main agent only (gated on agent_id); it re-fires on clear/compact, so the brief survives both. Requires jq.
</quick_reference>
<reference_index>
standard/PROMODE_MAIN_AGENT.md — the main-agent brief (copy into .claude/ exactly)standard/hooks/promode-main-context.sh — the SessionStart hook (copy into .claude/hooks/ exactly)references/main-agent-delivery.md — why the brief is hook-delivered, not in CLAUDE.mdreferences/agent-knowledge-wiki.md — the interlinked-graph model for durable agent knowledge
</reference_index><workflows_index> | Workflow | Purpose | |----------|---------| | install.md | Set up the promode hook + brief in a project | | update.md | Refresh the brief + hook to the latest version | | audit.md | Check the install is present and current | </workflows_index>
<success_criteria> A correctly-installed project has:
.claude/PROMODE_MAIN_AGENT.md — exact copy of standard/PROMODE_MAIN_AGENT.md.claude/hooks/promode-main-context.sh — exact copy of the standard hook, executable.claude/settings.json — a SessionStart hook entry (matchers startup|resume|clear|compact) running the hookCLAUDE.md present as the knowledge root (created if it was missing; existing content preserved; never holds the orchestration brief)jq available on PATH
</success_criteria>testing
Inspect a failed or stalled subagent's transcript compactly to recover. Use when a background subagent failed, stalled, or its <result> summary isn't enough to decide what to do next, and you need to look inside its run without loading the whole transcript into your context.
testing
Create, edit, improve, or audit AgentSkills. Use when creating a new skill from scratch or when asked to improve, review, audit, tidy up, or clean up an existing skill or SKILL.md file. Also use when editing or restructuring a skill directory (moving files to references/ or scripts/, removing stale content, validating against the AgentSkills spec). Triggers on phrases like "create a skill", "author a skill", "tidy up a skill", "improve this skill", "review the skill", "clean up the skill", "audit the skill".
testing
Host security hardening and risk-tolerance configuration for OpenClaw deployments. Use when a user asks for security audits, firewall/SSH/update hardening, risk posture, exposure review, OpenClaw cron scheduling for periodic checks, or version status checks on a machine running OpenClaw (laptop, workstation, Pi, VPS).
testing
Create, edit, improve, or audit AgentSkills. Use when creating a new skill from scratch or when asked to improve, review, audit, tidy up, or clean up an existing skill or SKILL.md file. Also use when editing or restructuring a skill directory (moving files to references/ or scripts/, removing stale content, validating against the AgentSkills spec). Triggers on phrases like "create a skill", "author a skill", "tidy up a skill", "improve this skill", "review the skill", "clean up the skill", "audit the skill".