decision-memo
Decision Memo
One page, always. Built for decisions that need a record, not decisions someone can make in their head.
When to use
Trigger on requests for decision documents, options analyses, recommendations, or choice briefs. Do not use for status reports or informational summaries.
Required inputs
Ask for any missing inputs in one batch. Do not draft without the first three.
- The decision to be made, stated as a question (required)
- Who makes the decision, by when (required)
- Current context: what changed, what is forcing the decision now (required)
- Options already considered (optional; if provided, use them; if not, generate 2-4)
- Constraints: budget, legal, political, capacity (optional but usually needed)
- Success criteria: how will Ane know later whether the decision was right (optional)
Structure
Produce a memo with these sections, in this order, each under the named H2. Target length: one page, 400-500 words.
Decision required
One sentence ending in a question mark. Name the decider.
Context
Three to five sentences. What changed, why it matters now, what happens if no decision is made. No backstory that predates the triggering change.
Options
Two to four. Never just two unless the user specifies binary. For each:
- Option label: short noun phrase
- What it is: one sentence
- Pros: 2-4 bullets, specific, no generic benefits
- Cons: 2-4 bullets, including the worst realistic case
- Reversibility: one of fully reversible, reversible with cost, effectively irreversible. Explain in one clause.
- Cost: money, time, political capital, or capacity, in the currency that matters most
Recommendation
One option, named. Two-sentence reasoning. No hedging. If the recommendation is weak, say the recommendation is weak and name the deciding factor.
Risks
Three to five. Not the cons from above — the second-order risks specific to the recommendation. For each: the risk, the probability (low/medium/high), the mitigation, the owner of the mitigation.
Next steps
Numbered list. Who does what, by when. Start with the decision itself: "Decide by [date], route via [channel]." End with the first action after the decision.
Open questions
Optional. Use only if there is information the decider needs that this memo cannot provide. Use the ⚠️ Data gap: format from CLAUDE.md.
Writing rules
Follow CLAUDE.md writing style strictly:
- Active voice. Subject + verb + object.
- Sentences under 25 words. Split with periods if longer.
- Specific verbs: "approve" not "consider approving".
- No filler: no "in order to", "it should be noted", "as per".
- No nominalizations: "decide" not "make a decision".
- No hedging unless uncertainty is real.
- No em-dashes.
- No rhetorical questions.
- State the main point first in every paragraph.
Tone
Candid, not deferential. If the recommendation is obvious, say so. If the options are all bad, name that. If the decision can be delayed without cost, recommend delay. The memo serves the decider, not comfort.
Limitations
Do not invent facts, figures, or prior decisions. If a fact is needed and not provided, flag it with ⚠️ Data gap:. Do not pad to reach a target length. A 250-word memo is fine if the decision is clean.
More from gasserane/personal-skills
journal-reflection
Guide Ane through structured reflection at end of day, end of week, after a decision, or after encountering new learning material. Use when the user says "journal", "reflect on today", "weekly review", "what did I learn", "after-action review", or references writing to the MELdigitalgarden vault. Produces Obsidian-compatible markdown; writes directly to the vault when filesystem MCP is connected, otherwise returns content for manual save.
28daily-brief
Produce a scannable morning briefing for Ane (Senior MEL/SRHR Specialist at IPPF). Use when the user asks for "daily brief", "morning briefing", "start of day", "what's on today", "plan my day", or similar. Assembles today's priority, calendar, overdue items, pending decisions, and an energy check. Operates on user-provided inputs by default; richer when calendar/email MCP servers are connected.
27mel-framework-citation
Enforce IPPF/UNFPA/UNAIDS evidence-and-rigour citation standard on MEL/SRHR output. Tier-aware on placement — Tier 1 working brief uses an Evidence base line at end of section; Tier 2 publication uses inline citations; Tier 1 / junior-MEL keeps framework names visible in prose AND uses an annotated Evidence base line. Use whenever Ane produces a theory of change, evaluation design, indicator set, donor report, or SRHR programme analysis. Injects current authoritative framework versions with author and year, flags outdated versions, and applies the data-gap protocol. Do not use for non-MEL work.
27evidence-synthesis
Conduct a rigorous rapid evidence assessment or systematic-lite literature review for MEL/SRHR questions. Use when Ane asks for "evidence review", "literature review", "evidence synthesis", "REA", "what does the evidence say", "what do we know about", or similar. Produces a structured brief with question framing, method, findings by theme, confidence grading, and implications for programme or evaluation design. Does not invent citations.
27vi
Vi — HR Specialist and Execution Orchestrator for MEL/SRHR work. Receives an approved plan from Ann (or directly from Ane), designs the specialist roster, spawns specialists as subagents, reviews their outputs, compiles the final product, and returns it. General-purpose — invoked by Ann via Agent tool, or directly by Ane when a plan is already approved.
26li
Li — Knowledge Manager for Ane's library and MEL Wiki. Use when Ane needs to catalog, retrieve, or reorganize documents in the personal knowledge library, or query/maintain the MEL Wiki. Handles INGEST, QUERY, and LINT operations. Does not answer domain questions — retrieves and organizes knowledge for other agents and Ane.
26