/founder-os:prompt:get

Retrieve a single prompt by name from the Team Prompt Library, collect values for any {{variable}} placeholders via interactive prompting, and output the fully substituted prompt ready to use.

Load Skills

Read the prompt-management skill at ${CLAUDE_PLUGIN_ROOT}/skills/prompt/prompt-management/SKILL.md before proceeding.

Parse Arguments

Extract from $ARGUMENTS:

• name (required positional): The name of the prompt to retrieve. This is all text provided before any -- flags. If no name is provided, ask the user: "Which prompt would you like to retrieve? Provide the prompt name." Then stop and wait for input.

Business Context (Optional)

Check if context files exist at _infrastructure/context/active/. If the directory contains .md files, read business-info.md, strategy.md, and current-data.md. Use this context to personalize output (e.g., prioritize known clients, use correct terminology, align with current strategy). If files don't exist, skip silently.

Step 0: Memory Context

Read the context-injection skill at _infrastructure/memory/context-injection/SKILL.md. Query for memories relevant to the current input (company, contacts, topics detected in arguments). If memories are returned, incorporate them into your working context for this execution.

Step 1: Verify Notion Availability

Check that the Notion MCP server is connected. If Notion is unavailable, display:

Error: Notion MCP is required for /founder-os:prompt:get.
See ${CLAUDE_PLUGIN_ROOT}/INSTALL.md for setup instructions.

Then stop.

Step 2: Locate the Team Prompt Library Database

Search the user's Notion workspace for a database named "[FOS] Prompts". If not found, try "Founder OS HQ - Prompts". If not found, fall back to "Team Prompt Library - Prompts".

If neither database exists, display:

Your Team Prompt Library is empty — no prompts have been added yet.

Get started with:
  /founder-os:prompt:add [prompt-text] --name="My First Prompt" --category=general

Then stop.

Step 3: Search for the Prompt by Name

Query the discovered prompts database for a page whose Name property matches the requested name. Apply case-insensitive matching.

If an exact match (case-insensitive) is found, proceed to Step 4.

If no exact match is found, search for partial matches: query for pages whose Name contains any word from the requested name. Collect up to 5 partial matches sorted by Times Used descending.

If partial matches are found, display:

Prompt not found: "[requested name]"

Did you mean one of these?

1. [Name] — [Category]  (used [N] times)
2. [Name] — [Category]  (used [N] times)
3. [Name] — [Category]  (used [N] times)

Run /founder-os:prompt:list to browse all prompts, or /founder-os:prompt:get "[exact name]" to retry.

Then stop.

If no matches at all are found (exact or partial), display:

Prompt not found: "[requested name]"

No similar prompts were found in your library. Check the name and try again.

Run /founder-os:prompt:list to browse all available prompts.

Then stop.

Step 4: Display Prompt Metadata

Display the prompt metadata before showing the content:

## [Name]

- **Category**: [Category]
- **Visibility**: [Personal / Shared]
- **Author**: [Author]
- **Tags**: [tag1, tag2, ...]  (or "None" if empty)
- **Times Used**: [N]
- **Last Used**: [YYYY-MM-DD]  (or "Never" if null)

Step 5: Detect Variables

Apply the variable detection regex from the prompt-management skill to the prompt's Content field:

\{\{([a-zA-Z_][a-zA-Z0-9_]*)\}\}

Extract all unique variable names (case-insensitive deduplication). Preserve the original casing of the first occurrence of each variable name for display purposes.

If no variables are detected, skip Step 6 and proceed directly to Step 7 with the prompt content as-is.

If variables are detected, display them before collecting values:

This prompt contains [N] variable(s) that need values:
  {{variable_1}}, {{variable_2}}, ...

Then proceed to Step 6.

Step 6: Collect Variable Values (only if variables were detected)

For each unique variable detected, use AskUserQuestion to collect a value. Ask one variable at a time in the order they first appear in the prompt content.

Prompt format for each variable:

Enter value for {{variable_name}}:

If a variable name uses a suffix convention, include a hint in parentheses:

• Variable ending in _date → hint: "(e.g., 2026-03-05)"
• Variable ending in _list → hint: "(comma-separated)"
• Variable ending in _tone or tone appears in the name → hint: "(e.g., formal, friendly, concise)"

Wait for the user's response before asking for the next variable. Do not ask for all variables in one message.

After all values are collected, perform substitution: replace every occurrence of {{variable_name}} in the prompt content with the user-supplied value. The replacement is case-insensitive — {{ClientName}} and {{clientname}} are treated as the same variable and receive the same value.

Step 7: Output the Substituted Prompt

Display the final, ready-to-use prompt in a fenced code block so it can be copied cleanly:

## Ready-to-Use Prompt

```
[full substituted prompt content here]
```

If no variables were present (Step 6 was skipped), display the prompt content unchanged in the same fenced format.

Below the code block, display a usage summary:

---

**Prompt**: [Name]
**Category**: [Category]
**Variables filled**: [N]  (or "None" if no variables)

Copy the prompt above and paste it into any AI tool or workflow.

Step 8: Update Usage Tracking

After displaying the prompt output, update the Notion page for the retrieved prompt:

1. Increment the Times Used counter by 1.
2. Set Last Used to the current date in YYYY-MM-DD format.

Perform both updates in a single Notion page update call. Do not skip tracking — a retrieval is always a usage event, regardless of whether variables were filled.

Final: Memory Update

Read the pattern-detection skill at _infrastructure/memory/pattern-detection/SKILL.md. Log this execution as an observation with: plugin name, primary action performed, key entities (companies, contacts), and output summary. Check for emerging patterns per the detection rules. If a memory reaches the adaptation threshold, append the notification to the output.

Graceful Degradation

• Notion unavailable: Show error message with INSTALL.md reference and stop (see Step 1).
• Database missing: Show empty library message with onboarding hint and stop (see Step 2).
• Prompt not found (exact): Show similar names if partial matches exist, otherwise show not-found message (see Step 3).
• Variable collection interrupted: If the user provides an empty value for a variable, use the literal placeholder {{variable_name}} as the substituted value and continue — do not block on empty input.
• Usage tracking fails: If the Notion update in Step 8 fails, do not surface the error to the user. The prompt output has already been displayed and that is the primary outcome.
• Content exceeds 2000 characters: If the prompt content is split across multiple rich_text blocks in Notion, concatenate all blocks in order before running variable detection and substitution.

Usage Examples

/founder-os:prompt:get "Client Intro Email"
/founder-os:prompt:get "Weekly Status Update"
/founder-os:prompt:get cold outreach
/founder-os:prompt:get "Code Review Checklist"
/founder-os:prompt:get competitor analysis