How to Force-Load a Specific PM Skills Skill: Complete Syntax Guide
Prefix the skill name with a forward slash using either /plugin-name:skill-name or /skill-name to force-load a PM Skills skill immediately without waiting for trigger phrase matching.
The phuryn/pm-skills repository provides a collection of product management skills for Claude that typically activate automatically based on conversation context. When you need to prioritize specific domain knowledge over general model responses, you can force-load a skill using explicit slash commands documented in the project's README.md and CLAUDE.md files.
Understanding Auto-Loading vs Force-Loading
PM Skills are built as individual markdown files containing structured knowledge and trigger phrases. According to the repository's README.md, "Skills are loaded automatically when relevant to the conversation — no explicit invocation needed." This auto-loading mechanism watches for semantic matches between user utterances and skill trigger phrases.
However, certain scenarios require immediate skill availability. As noted in CLAUDE.md, you can bypass the trigger-matching phase to ensure a skill's knowledge base is immediately injected into the conversation context.
Exact Syntax for Force-Loading
The force-load syntax supports two formats depending on whether you specify the plugin owner.
Full Plugin-Skill Syntax
Use the colon-separated format when you need to specify exactly which plugin owns the skill:
/plugin-name:skill-name
This format is unambiguous and works regardless of which plugins are installed. For example, loading the resume review skill from the pm-toolkit plugin:
/pm-toolkit:review-resume
Short-Form Skill-Only Syntax
You can also use the abbreviated format:
/skill-name
When using the short form, Claude automatically resolves the plugin prefix if the skill name is unique across installed plugins. As documented in README.md, "Claude will add the prefix" automatically when you use /skill-name.
Practical Force-Loading Examples
Resume Review Skill (pm-toolkit)
To immediately access the resume review capabilities documented in pm-toolkit/skills/review-resume/SKILL.md:
User: /pm-toolkit:review-resume
Claude: (skill `review-resume` is now loaded)
User: Here is my resume PDF…
Using the short form achieves the same result:
User: /review-resume
Claude: (skill `review-resume` loads automatically)
User: Please tailor this resume for a senior PM role.
Strategy Red Team Skill (pm-execution)
For the strategy assessment skill located at pm-execution/skills/strategy-red-team/SKILL.md:
User: /pm-execution:strategy-red-team
Claude: (skill `strategy-red-team` is now active)
User: Analyze the assumptions in this product roadmap.
CLI Usage with Claude Code
When using Claude Code or Codex CLI, first install the plugin containing the target skill:
claude plugin install pm-toolkit@pm-skills
Then force-load it in an interactive session:
claude> /pm-toolkit:review-resume
How Force-Loading Works Internally
The mechanism operates through three distinct phases:
- Skill Registration: Each skill resides under
plugin-name/skills/and contains front-matter with adescriptionand trigger phrases. - Auto-Loading: Claude monitors user utterances for trigger phrase matches before fetching the skill content.
- Force-Loading: The forward-slash prefix signals Claude to load the skill immediately without waiting for semantic matching, ensuring the skill's knowledge becomes part of the model's context for the next prompt.
Summary
- Force-load syntax: Use
/plugin-name:skill-namefor explicit plugin specification or/skill-namefor automatic resolution. - Immediate availability: Force-loading bypasses trigger phrase matching to inject skill knowledge instantly.
- Key files: Documentation in
README.mdandCLAUDE.mdconfirms both syntax variants are officially supported. - Use cases: Prioritize specific PM domain knowledge over general model responses when analyzing resumes, roadmaps, or strategy documents.
Frequently Asked Questions
What is the difference between auto-loading and force-loading PM Skills?
Auto-loading activates skills automatically when Claude detects trigger phrases in your conversation that match the skill's registered keywords. Force-loading uses the /skill-name or /plugin-name:skill-name syntax to immediately inject a skill's knowledge into the context without waiting for semantic triggers, ensuring the specific expertise is available for your next prompt.
Do I need to install the plugin before force-loading a skill?
Yes, the plugin containing the skill must be installed before force-loading will work. As shown in the CLI examples, you must first run claude plugin install pm-toolkit@pm-skills (or the appropriate plugin name) before you can force-load any skills contained within that plugin.
Can I force-load multiple PM Skills simultaneously?
While the documentation focuses on single skill loading, you can issue multiple force-load commands in sequence. Each /plugin-name:skill-name command loads that specific skill's content into the conversation context. There is no explicit syntax shown for loading multiple skills in a single command, so issue separate commands for each skill you need active.
Why does Claude sometimes add the plugin prefix automatically?
When you use the short form /skill-name, Claude resolves the full plugin path automatically if the skill name is unique across your installed plugins. This behavior is documented in the README.md quote: "Claude will add the prefix." If multiple plugins contain skills with identical names, you must use the full /plugin-name:skill-name syntax to specify which implementation to load.
Have a question about this repo?
These articles cover the highlights, but your codebase questions are specific. Give your agent direct access to the source. Share this with your agent to get started:
curl -s "https://instagit.com/install.md" Maintain an open-source project? Get it listed too →