@human-avatar/skills-for-humanity 1.0.0

package/skills/systems-leverage-analysis/SKILL.md

@@ -0,0 +1,65 @@
+---
+name: systems-leverage-analysis
+description: "Finds where small interventions produce large, lasting change using Donella Meadows' leverage point hierarchy. Use when asked 'where should we intervene', 'highest leverage', 'what actually changes this system', or 'find the lever'."
+---
+
+# Systems Leverage Analysis
+
+Most interventions target low-leverage parameters — adjusting numbers, tweaking rates — when high-leverage structural points are available and being ignored. Donella Meadows identified 12 places to intervene in a system, ranging from parameters (nearly powerless) to paradigm (most powerful). The reason high-leverage points go unused is that they face the highest resistance; understanding this is part of the analysis.
+
+---
+
+## Your Process
+
+**Step 1: List Candidate Interventions**
+Gather all interventions currently being considered or tried. Include past attempts that failed.
+
+**Step 2: Classify by Meadows Hierarchy**
+Map each intervention to its leverage level:
+- **Low:** numbers/parameters, buffer sizes, flow rates
+- **Medium:** feedback loop strength, information flows, rules of the system
+- **High:** goals of the system, system structure, paradigm (the beliefs that create the system)
+
+**Step 3: Identify the Default Level**
+What level is typically targeted — and why? Understand the political, cognitive, or practical reasons low-leverage points get chosen.
+
+**Step 4: Surface Ignored High-Leverage Points**
+What higher-leverage interventions exist that are not being tried? Trace why they are being avoided (too costly, politically threatening, requires belief change, long time horizon).
+
+**Step 5: Assess Feasibility**
+High-leverage points often face disproportionate resistance. For each high-leverage option: what would be required to act on it? Is it feasible given current constraints?
+
+---
+
+## Human Check-in
+
+Before proceeding, ask the user:
+
+**How do you want to run this?**
+
+- **A) Full analysis** — complete all steps, reasoning shown throughout
+- **B) Key findings only** — bottom-line output, skip step-by-step detail
+- **C) Highest leverage point only** — single best intervention, skip lower-leverage options
+- **D) Refine the framing** — adjust what we're analyzing before starting
+
+Proceed based on their choice.
+
+## Output Format
+
+**Intervention Table**
+
+| Intervention | Leverage Level | Leverage Type | Feasibility | Resistance Source |
+|-------------|---------------|--------------|-------------|-------------------|
+| | | | | |
+
+**Default Level Being Targeted:** [level + reason]
+
+**Highest-Leverage Feasible Point:** [intervention + why it's higher leverage + what unlocks it]
+
+**Ignored High-Leverage Options:** [what they are + why they're being avoided]
+
+---
+
+## Notes
+
+High-leverage points are often counterintuitive — pushing harder in the obvious direction can make things worse. Identify any points where the intuitive intervention is actually negative leverage.

package/skills/temporal/SKILL.md

@@ -0,0 +1,78 @@
+---
+name: temporal
+description: "Entry point for the temporal thinking toolkit. Routes to the right temporal skill based on your situation. Use when you say 'temporal', 'timing', 'when', 'long term', 'what happens in 5 years', 'what cycle is this', 'should we wait', 'possible futures', or want time-horizon thinking applied without knowing which specific tool fits."
+---
+
+# Temporal
+
+Applies temporal reasoning to timing, horizons, cycles, and futures. Diagnoses what kind of time-based analysis is needed and applies the right tool.
+
+## Which tool fits
+
+| You need to... | Tool |
+|---|---|
+| Identify what cycle this is and where in it you are | cycle-detection |
+| Map consequences across short, medium, and long horizons | horizon-mapping |
+| Explore possible, probable, and preferable futures | futures-mapping |
+| Assess whether now is the right time to act | timing-analysis |
+
+## Routing Decision
+
+- **Situation feels like a pattern that repeats — want to know where in it** → cycle-detection
+- **Decision has different implications at different time scales** → horizon-mapping
+- **Planning for an uncertain future with multiple possible paths** → futures-mapping
+- **Wondering whether to act now or wait** → timing-analysis
+- **Unclear** → horizon-mapping; mapping consequences across time scales usually surfaces timing questions, cycles, and futures simultaneously
+
+## Confirm Direction
+
+After diagnosing which tool fits, present the recommendation before executing:
+
+> My read: **[diagnosed tool]** — one sentence on why it fits.
+
+- **A) Yes, run that tool**
+- **B) Show me all options** — list every skill in this category with one-line descriptions
+- **C) Quick version** — lighter-weight alternative for this situation, if one exists
+- **D) Re-diagnose** — describe the situation differently for a second read
+
+Wait for their selection before proceeding.
+
+---
+
+## Cycle Detection
+
+*Identifies what recurring cycle this is and where you currently are.*
+
+Most situations are instances of known cycles: hype cycles, market cycles, innovation S-curves, organizational life cycles, political pendulums. Name the cycle candidates. Test each: does the current situation match the structural characteristics of this cycle? What is the evidence for each phase position? Knowing where in the cycle you are changes what the right action is — the same move is correct in one phase and disastrous in another.
+
+**Output:** Cycle identified with evidence, current phase, what the cycle predicts comes next, and what the phase position implies for current decisions.
+
+---
+
+## Horizon Mapping
+
+*Maps consequences of a decision across short, medium, and long time horizons.*
+
+Many decisions that look good in the short term are harmful in the long term, and vice versa. Map the consequences of each option at three horizons: (1) Short — what changes in the next 1-3 months? (2) Medium — what does this look like in 1-2 years? (3) Long — what is the 5-10 year trajectory? Look especially for short-term gains that create long-term lock-in or erosion, and short-term costs that build long-term capacity.
+
+**Output:** Three-horizon consequence map for each option, with particular attention to short-term/long-term trade-offs and lock-in risks.
+
+---
+
+## Futures Mapping
+
+*Explores possible, probable, and preferable futures.*
+
+Use the futures cone: (1) Possible — what could happen (full range), (2) Plausible — what could reasonably happen given current trends, (3) Probable — what is likely to happen if current trajectories continue, (4) Preferable — what should happen and how to move toward it. For each quadrant: identify the key uncertainties driving divergence. Build 3 distinct scenarios: the continuation, the disruption, and the transformation.
+
+**Output:** Futures map across all four quadrants, key uncertainties, three scenarios (continuation/disruption/transformation), and what this implies for decisions being made today.
+
+---
+
+## Timing Analysis
+
+*Assesses whether now is the right time to act, wait, or prepare.*
+
+Timing matters as much as the action itself. Assess four dimensions: (1) Readiness — is the capability to execute actually present? (2) Window — is there a time-limited opportunity or threat? (3) Momentum — is the context moving toward or away from favorable conditions? (4) Cost of waiting — does delay make things harder, or allow important information to emerge? The right action at the wrong time often fails; the same action at the right time succeeds.
+
+**Output:** Assessment across all four dimensions, the timing recommendation (act now / wait / prepare to act), and the conditions that would trigger reassessment.

package/skills/temporal-cycle-detection/SKILL.md

@@ -0,0 +1,75 @@
+---
+name: temporal-cycle-detection
+description: "Identifies what recurring cycle a situation is an instance of and where in that cycle you currently are. Use when asked 'what cycle is this', 'where are we in the cycle', 'have we seen this before', 'detect the pattern', 'hype cycle', or 'where in the curve'."
+---
+
+# Temporal Cycle Detection
+
+Every situation that feels unprecedented is usually an instance of a recurring cycle. Knowing your position in the cycle tells you what phase is coming, what actions are appropriate now versus premature, and which signals indicate divergence from the typical pattern — the most important signal of all. Acting as if a situation is unique when it is not forfeits the pattern's predictive value.
+
+---
+
+## Your Process
+
+**Step 1: Describe Current State and Trajectory**
+State what is happening now and how the situation has moved over the past period. Include: rate of change, sentiment, who is entering or exiting, resource flows, expectations.
+
+**Step 2: Match to a Candidate Cycle**
+Compare against the most common cycles:
+- **Technology Hype Cycle (Gartner):** trigger → peak of inflated expectations → trough of disillusionment → slope of enlightenment → plateau of productivity
+- **Product Adoption Curve (Rogers):** innovators → early adopters → early majority → late majority → laggards
+- **Business/Economic Cycle:** expansion → peak → contraction → trough → recovery
+- **Organisational Change Curve (Kübler-Ross adapted):** shock → denial → frustration → depression → experiment → decision → integration
+- **Competitive Cycle:** emergence → growth → shakeout → maturity → decline/renewal
+- **Market Cycle:** accumulation → markup → distribution → markdown
+
+**Step 3: Map Current Position**
+Place the situation on the curve. Be specific about which phase and where within the phase (early, mid, late).
+
+**Step 4: Identify Phase Signs**
+List the characteristic indicators of this phase. Which are present? Which are absent? Absent expected signs are as informative as present ones.
+
+**Step 5: State the Typical Next Phase**
+What normally follows this phase? What are the leading indicators that the transition has begun?
+
+**Step 6: Note Divergences**
+Where does this situation differ from the typical cycle? These divergences — not the pattern itself — are the most important signals. They indicate either a different cycle applies or this instance will unfold differently.
+
+---
+
+## Human Check-in
+
+Before proceeding, ask the user:
+
+**How do you want to run this?**
+
+- **A) Full analysis** — complete all steps, reasoning shown throughout
+- **B) Key findings only** — bottom-line output, skip step-by-step detail
+- **C) Current position only** — where in the cycle we are right now, skip the full cycle description
+- **D) Refine the framing** — adjust what we're analyzing before starting
+
+Proceed based on their choice.
+
+## Output Format
+
+**Cycle Match:** [cycle name] — [why it fits]
+
+**Current Position:** [phase name] — [early / mid / late within phase]
+
+**Phase Signs**
+
+| Expected Sign for This Phase | Present? |
+|------------------------------|---------|
+| | |
+
+**Typical Next Phase:** [name + leading indicators of transition]
+
+**Divergences:** [where this instance differs from the typical pattern + what that implies]
+
+**Implications:** [what actions are appropriate now vs. premature vs. overdue]
+
+---
+
+## Notes
+
+If no cycle fits well, that is itself a finding — either the situation is genuinely novel or it requires a composite of two cycles. Name the mismatch explicitly rather than forcing a fit.

package/skills/temporal-futures-mapping/SKILL.md

@@ -0,0 +1,63 @@
+---
+name: temporal-futures-mapping
+description: "Explores possible, probable, and preferable futures using scenario thinking. Use when asked about 'scenario planning', 'futures thinking', 'possible futures', 'how might this play out', 'futures cone', or 'what could happen'."
+---
+
+# Temporal Futures Mapping
+
+The future is not a single path — it is a cone of possibilities that narrows as time passes and choices are made. Scenario thinking does not predict the future; it maps the cone so that plans can be tested against multiple plausible worlds rather than a single assumed one. Plans that only work in one scenario are fragile. Plans robust across several are resilient.
+
+---
+
+## Your Process
+
+**Step 1: Define Time Horizon and Decision Context**
+State the specific decision or question being stress-tested and the time horizon (1 year, 5 years, 10 years). The scenarios must be built around a decision, not just as general futures.
+
+**Step 2: Identify Key Uncertainties**
+Find the 2–3 variables that most shape outcomes but are least predictable. These are the axes along which scenarios diverge. Avoid certainties (they are part of all scenarios) and avoid trivialities (they don't change much).
+
+**Step 3: Build 3–4 Distinct Scenarios**
+Span the cone — from possible to plausible to probable to preferable. Each scenario should be:
+- Named (a vivid name aids recall)
+- Described (the world in this future — what is true, who won, what changed)
+- Traceable (the path from now to that world — the sequence of events that produced it)
+
+**Step 4: Test Current Plans**
+For each scenario: does the current plan work? Does it fail? Does it create new problems? Plans that are robust across all scenarios are high-confidence. Plans that only work in the most optimistic scenario require hedging.
+
+---
+
+## Human Check-in
+
+Before proceeding, ask the user:
+
+**How do you want to run this?**
+
+- **A) Full analysis** — complete all steps, reasoning shown throughout
+- **B) Key findings only** — bottom-line output, skip step-by-step detail
+- **C) Most probable future only** — skip possible and preferable variants, focus on what's likely
+- **D) Refine the framing** — adjust what we're analyzing before starting
+
+Proceed based on their choice.
+
+## Output Format
+
+**Key Uncertainties:** [the 2–3 variables driving divergence, with their plausible ranges]
+
+**Scenario Table**
+
+| Name | Description | Path to This World | Implications for Current Plan |
+|------|-------------|-------------------|------------------------------|
+| | | | |
+
+**Robustness Assessment**
+- Robust across all scenarios: [what holds]
+- Fragile (only works in one scenario): [what is at risk]
+- Hedging required: [where to build optionality]
+
+---
+
+## Notes
+
+Scenarios are not predictions — they are structured hypotheses. The goal is not to determine which will occur but to identify which actions are robust regardless of which occurs.

package/skills/temporal-horizon-mapping/SKILL.md

@@ -0,0 +1,65 @@
+---
+name: temporal-horizon-mapping
+description: "Maps consequences of a decision across short, medium, and long time horizons. Use when asked about 'time horizons', 'what does this look like in 5 years', 'short-term vs long-term', 'map the implications over time', or 'think through the long-term'."
+---
+
+# Temporal Horizon Mapping
+
+Decisions that look good now often look very different at 1, 3, or 10 years. The most consequential errors in judgment come not from bad reasoning in the moment but from evaluating a decision at the wrong time horizon — optimizing for the immediate while the real costs land later. Making all three horizons explicit forces the tradeoff into view rather than leaving it implicit.
+
+---
+
+## Your Process
+
+**Step 1: State the Decision**
+Name the decision being evaluated and the current context in which it is being made. Clarity here prevents analysis drifting to adjacent decisions.
+
+**Step 2: Map Immediate Consequences (0–3 months)**
+What is the likely state immediately after acting? What resources are committed or freed? Who is affected and how? What has this enabled or closed off in the near term?
+
+**Step 3: Map Medium-Term Consequences (6–24 months)**
+What does the situation look like after the initial effects have compounded? What second-order effects emerge? Who gains or loses standing? What dependencies or path-dependencies have formed?
+
+**Step 4: Map Long-Term Consequences (3+ years)**
+What has the decision made likely or unlikely at scale and over time? What is the structural change — to capabilities, relationships, markets, culture? What would be very difficult to reverse by this point?
+
+**Step 5: Flag Reversals**
+Identify decisions that look positive short-term but create long-term problems — and the reverse. These reversals are the highest-value output of this analysis.
+
+**Step 6: Identify the Governing Horizon**
+At which horizon do the most significant consequences actually land? Is that the horizon currently being used to evaluate this decision? Mismatched horizons are the primary source of poor long-term decisions made in good faith.
+
+---
+
+## Human Check-in
+
+Before proceeding, ask the user:
+
+**How do you want to run this?**
+
+- **A) Full analysis** — complete all steps, reasoning shown throughout
+- **B) Key findings only** — bottom-line output, skip step-by-step detail
+- **C) Long-term consequences only** — what emerges beyond the obvious timeframe, skip near-term
+- **D) Refine the framing** — adjust what we're analyzing before starting
+
+Proceed based on their choice.
+
+## Output Format
+
+**Horizon Table**
+
+| Horizon | Timeframe | Likely State | Enabled / Foreclosed | Who Is Better / Worse Off |
+|---------|----------|-------------|---------------------|--------------------------|
+| Immediate | 0–3 months | | | |
+| Medium | 6–24 months | | | |
+| Long | 3+ years | | | |
+
+**Reversal Flags:** [decisions that look good short-term but create long-term problems, or vice versa]
+
+**Governing Horizon:** [which horizon should drive this decision + why it differs from current evaluation horizon if applicable]
+
+---
+
+## Notes
+
+Short-term and long-term are not automatically in conflict — some decisions improve all horizons. Identify those as high-confidence choices; the real analysis is needed where horizons diverge.

package/skills/temporal-timing-analysis/SKILL.md

@@ -0,0 +1,67 @@
+---
+name: temporal-timing-analysis
+description: "Assesses whether now is the right time to act, wait, or prepare. Use when asked 'is now the right time', 'should we wait', 'are we too early', 'timing this decision', 'when to move', or 'momentum'."
+---
+
+# Temporal Timing Analysis
+
+Timing is often as consequential as the decision itself. The right action at the wrong time fails — too early before conditions are ready, too late after the window has closed. Most timing judgments are made implicitly, with the urgency of the moment substituting for analysis. Making timing explicit means identifying which conditions are present, which are absent, and whether the environment is moving toward or away from readiness.
+
+---
+
+## Your Process
+
+**Step 1: State the Action and Current Context**
+Name the action under consideration and describe current conditions — market, organizational, political, technical. The timing analysis is grounded in this specific context.
+
+**Step 2: Identify Readiness Conditions**
+What observable conditions would make this optimal timing? These are specific and external — not "when we're ready" but "when X is true in the environment". For each condition: is it currently present or absent?
+
+**Step 3: Assess Momentum**
+Is the environment moving toward or away from ideal conditions? A missing condition that is approaching matters differently from one that is receding. Momentum changes the urgency calculation.
+
+**Step 4: Cost of Waiting**
+What is lost or foregone per unit of delay? Is the window closing — and how fast? Are competitors moving? Is the opportunity time-limited? Quantify where possible.
+
+**Step 5: Cost of Acting Early**
+What risks come from moving before conditions are right? First-mover disadvantage, resource waste, organizational fatigue from premature initiatives, credibility cost of a failed early attempt.
+
+**Step 6: Define Trigger Signals**
+What specific observable events should prompt action? These make timing a decision rule rather than a repeated judgment call.
+
+---
+
+## Human Check-in
+
+Before proceeding, ask the user:
+
+**How do you want to run this?**
+
+- **A) Full analysis** — complete all steps, reasoning shown throughout
+- **B) Key findings only** — bottom-line output, skip step-by-step detail
+- **C) Timing verdict only** — act, wait, or prepare — and the single most important reason why
+- **D) Refine the framing** — adjust what we're analyzing before starting
+
+Proceed based on their choice.
+
+## Output Format
+
+**Readiness Conditions**
+
+| Condition | Present? | Momentum (toward / away) |
+|-----------|---------|--------------------------|
+| | | |
+
+**Cost of Waiting:** [what is lost per period of delay; is the window closing?]
+
+**Cost of Acting Early:** [risks of premature action]
+
+**Timing Recommendation:** [act now / wait / prepare now, act at trigger] + rationale
+
+**Trigger Signals:** [specific observable events that should prompt action]
+
+---
+
+## Notes
+
+If both costs (waiting and acting early) are high, the situation requires a staged approach — begin preparation now, commit fully at the trigger. Identify what preparation can happen safely before the trigger is reached.

package/skills/writing/SKILL.md

@@ -0,0 +1,115 @@
+---
+name: writing
+description: "Routes to the right writing skill for any fiction, non-fiction, or professional writing challenge. Use when you say 'my character feels flat', 'the story isn't working', 'this scene falls flat', 'the dialogue sounds wrong', 'the prose is clunky', 'the argument isn't landing', 'write a report', 'executive summary', 'the tone keeps shifting', 'line editing', 'my world feels thin', 'POV problems', 'inconsistency check', 'write copy', 'technical documentation', 'audience calibration', 'rhetorical analysis', or any time you have a writing problem but aren't sure which specific tool fits."
+---
+
+# Writing
+
+Every writing problem has a name. Name it precisely, and the fix becomes obvious. The most common mistake is treating a structural problem as a prose problem, or a character problem as a plot problem — applying the wrong tool while the real issue goes unaddressed.
+
+This routing skill reads your situation, identifies the problem type, and connects you to the right technique immediately.
+
+---
+
+## Your Process
+
+### Step 1: Read the Situation
+
+If the user hasn't described their writing challenge yet, ask:
+
+> What's the writing problem? Paste the relevant passage, describe the situation, or tell me what isn't working.
+
+Wait for their response before diagnosing.
+
+---
+
+### Step 2: Diagnose the Problem Type
+
+Read the situation and classify it against these problem clusters:
+
+**Fiction craft** — The core problem is in the storytelling machinery: character psychology, plot architecture, scene structure, dialogue texture, world credibility, narrative arc, or point-of-view integrity.
+→ Route to: `/writing-character-development`, `/writing-plot-structure`, `/writing-scene-construction`, `/writing-dialogue`, `/writing-worldbuilding`, `/writing-arc-design`, `/writing-pov`
+
+**Continuity and consistency** — Something in the story contradicts itself: timeline errors, character knowledge violations, world-rule inconsistencies, physical continuity breaks.
+→ Route to: `/writing-inconsistency-audit`
+
+**Editing and revision** — The material exists but needs structural reordering, tonal coherence, sentence-level cleaning, prose quality elevation, or voice stabilisation.
+→ Route to: `/writing-restructure`, `/writing-tone-alignment`, `/writing-line-editing`, `/writing-prose-elevation`, `/writing-voice-consistency`
+
+**Professional and non-fiction** — A real-world document needs to be written or improved: a business report, marketing copy, technical documentation, analytical argument, or executive brief.
+→ Route to: `/writing-report`, `/writing-copy`, `/writing-technical`, `/writing-argument`, `/writing-executive-summary`
+
+**Rhetoric and audience** — The writing needs to be calibrated for a specific reader, or the rhetorical moves in a piece need to be surfaced and examined.
+→ Route to: `/writing-audience-calibration`, `/writing-rhetoric`
+
+---
+
+### Step 3: Identify 3–4 Best-Fit Skills
+
+Within the most relevant cluster, identify the 2–4 skills that best match the specific problem. Use the routing guide below.
+
+Prioritize skills that match the **type of output** the user needs:
+- They need a diagnosis → audit skills (`/writing-inconsistency-audit`, `/writing-pov`, `/writing-scene-construction`)
+- They need a rebuild → design skills (`/writing-plot-structure`, `/writing-arc-design`, `/writing-character-development`)
+- They need a revision → editing skills (`/writing-restructure`, `/writing-line-editing`, `/writing-tone-alignment`)
+- They need a produced piece → production skills (`/writing-report`, `/writing-executive-summary`, `/writing-copy`)
+- They need a rhetorical read → analysis skills (`/writing-rhetoric`, `/writing-argument`)
+
+---
+
+### Step 4: Present Options
+
+Present the diagnosis and options clearly:
+
+> Here's what I think you need. Which fits your situation best?
+>
+> **A) [Skill name]** — [one sentence on why this fits and what it produces]
+> **B) [Skill name]** — [one sentence on why this fits and what it produces]
+> **C) [Skill name]** — [one sentence on why this fits and what it produces]
+> **D) Show me more options** — list all skills in the Writing category
+
+Wait for their selection.
+
+---
+
+### Step 5: Execute
+
+When the user picks an option:
+- **A, B, or C:** Run the selected skill immediately. Do not ask them to type another command. Use the context already gathered as the input.
+- **D:** Show the full skill table for the Writing category. Let them pick from the complete list, then execute.
+
+---
+
+## Routing Guide
+
+| Situation | Top skills to offer |
+|---|---|
+| "My character feels flat / doesn't feel real" | `/writing-character-development` |
+| "The story isn't working / the plot sags" | `/writing-plot-structure`, `/writing-arc-design` |
+| "This scene isn't working / feels flat" | `/writing-scene-construction`, `/writing-dialogue` |
+| "The dialogue sounds wrong / on the nose" | `/writing-dialogue`, `/writing-character-development` |
+| "The world feels thin / like a backdrop" | `/writing-worldbuilding` |
+| "The arc feels unearned / the ending doesn't land" | `/writing-arc-design`, `/writing-plot-structure` |
+| "POV problems / head-hopping / narration inconsistent" | `/writing-pov`, `/writing-inconsistency-audit` |
+| "Contradictions / continuity errors" | `/writing-inconsistency-audit` |
+| "The piece is in the wrong order / buries the lede" | `/writing-restructure`, `/writing-executive-summary` |
+| "The tone keeps shifting / voice drift" | `/writing-tone-alignment`, `/writing-voice-consistency` |
+| "The sentences are clunky / wordy / passive" | `/writing-line-editing` |
+| "The prose is flat / competent but not compelling" | `/writing-prose-elevation` |
+| "Different contributors / brand voice inconsistent" | `/writing-voice-consistency` |
+| "Write or fix a report / briefing document" | `/writing-report`, `/writing-executive-summary` |
+| "Marketing copy / landing page / ad copy" | `/writing-copy`, `/writing-audience-calibration` |
+| "Technical documentation / API docs / user guide" | `/writing-technical`, `/writing-audience-calibration` |
+| "Build an argument / op-ed / make the case" | `/writing-argument`, `/writing-rhetoric` |
+| "Executive summary / 1-page brief for leadership" | `/writing-executive-summary`, `/writing-report` |
+| "Write for a specific audience / calibrate" | `/writing-audience-calibration` |
+| "Rhetorical analysis / what is this piece doing" | `/writing-rhetoric`, `/writing-argument` |
+
+---
+
+## Notes
+
+- **Don't confuse problem types.** A flat character is almost never a prose problem — it's a character-design problem. Line-editing flat prose won't fix it. Diagnose the layer first.
+- **Structure before style.** If the piece has structural problems, fix those before applying prose elevation or line editing. Rewriting sentences in the wrong order is wasted work.
+- **The context gathered in Step 1 is the input.** When the user selects a skill, run it on that context — they shouldn't have to re-explain the situation.
+- **If the situation spans multiple problem types** (e.g., a scene that has both dialogue problems and structural problems), present both skills and let them choose the highest-leverage starting point.

package/skills/writing-arc-design/SKILL.md

@@ -0,0 +1,68 @@
+---
+name: writing-arc-design
+description: "Maps and repairs character arcs and thematic arcs by aligning external plot events with internal change. Use when an arc feels unearned, a change feels sudden, or the ending doesn't land. Triggers: 'the arc feels unearned', 'arc design', 'the ending doesn't land', 'character arc', 'thematic arc', 'the change feels sudden', 'the character just changes at the end'."
+---
+
+# Writing: Arc Design
+
+Arcs feel unearned when the external plot and internal change are independent — when the character changes because the story requires it, not because the events forced it. This is the most common arc failure, and it produces the most visceral reader response: "the character just suddenly becomes a different person." The reader is right. The change happened on a schedule, not as a consequence.
+
+The fundamental principle: the external must create the conditions for the internal. The plot does not just happen *alongside* the character — it *acts on* the character in ways that make the internal change not just possible but necessary. The protagonist's wound must be pressed on, repeatedly, by the specific events of the plot. The defence must be made untenable by the story's demands. The arc endpoint is not the character deciding to change — it is the character having no remaining option except to change, or to break.
+
+The ending is not the change. The ending is the consequence of the change. This distinction matters enormously: if the character's internal shift and the external resolution happen simultaneously, the causation is lost. The internal shift must precede the external resolution, making it possible. The reader needs to see the new belief or new capacity tested against the story's central problem — and the test must require exactly what the old self couldn't have provided.
+
+---
+
+## Your Process
+
+**Step 1: Start State**
+What is the character's belief, wound, and worldview at the story's opening? Be specific: not "she is closed off" but "she believes that needing people is weakness — a belief installed by her mother's abandonment when she was nine, which she has managed by becoming self-sufficient to a degree that reads as coldness." The start state must be specific enough to be testable: you should be able to ask of any scene "is this character acting from their opening belief?" and get a clear answer.
+
+**Step 2: End State**
+What has changed? What is the new belief? Again, be specific. The end state should be the direct answer to the wound — not its opposite (which would be naive), but the more complex truth that the wound prevented the character from seeing. "She believes needing people is weakness" → "She learns that need is not weakness if the person chosen to be needed is trustworthy." The end state is not a lesson or a moral — it is a lived understanding, demonstrated in action.
+
+**Step 3: Pressure Points**
+Map the specific events that force the internal change. These must be caused by the plot — they are not random events that happen to push buttons; they are the story's central events landing on the character's specific vulnerability. For each pressure point: what does it press? What does the character's defence do in response? Does the defence hold, or does it start to crack?
+
+The arc requires at least three pressure points, and they should escalate: the first press is deflected, the second is harder to deflect, the third (the dark night) cannot be deflected at all.
+
+**Step 4: External-Internal Alignment**
+Map the parallel tracks:
+- External: [inciting incident → complication → midpoint shift → dark night → climax]
+- Internal: [defence holds → defence strained → defence fails → wound exposed → new belief tested]
+
+Do the external events directly generate the internal states? Or are they running on parallel tracks that don't connect? The test: could a different character, with a different wound and different defence, move through the same external events and arrive at the same ending? If yes, the plot is not doing character work — it is plot for its own sake.
+
+**Step 5: Thematic Argument**
+What does the story argue about the world? Not a message or a moral — an argument: a claim about how things are, how they should be, or what the cost of a particular way of living is. The ending should resolve this argument — not necessarily answer it definitively, but test it at full strength and show what the story believes.
+
+---
+
+## Output Format
+
+### Arc Map
+
+**Start State:** [Belief / wound / worldview — specific and testable]
+
+**End State:** [New belief — the wound's answer, not its opposite / how it is demonstrated in action]
+
+**Pressure Points:**
+1. [Event → what it presses → character response → does defence hold?]
+2. [Event → what it presses → character response → does defence crack?]
+3. [Event → what it presses → character response → defence fails]
+
+**External-Internal Alignment:** [Map of parallel tracks + diagnosis of where they connect or diverge]
+
+**Thematic Argument:** [What does the story claim? / Does the ending resolve it at full strength?]
+
+**Arc Verdict:** [Earned / Unearned / Partially earned — specific diagnosis of what is missing]
+
+---
+
+## Notes
+
+- The change must be earned by the pressure points, and the pressure points must be caused by the plot. Any link in this chain that is missing produces an unearned arc.
+- Slow arcs are not unearned arcs. A character who changes gradually across a long work can be completely earned if each pressure point genuinely presses and the defence genuinely fails. The problem is not pace — it is causation.
+- Pairs with `/writing-plot-structure` — the external structure is the mechanism that generates the pressure points; if the plot structure is broken, the arc cannot be earned.
+- Pairs with `/writing-character-development` — the arc endpoint must flow from the wound and defence; you cannot design the arc without the character profile.
+- Pairs with `/writing-scene-construction` when individual scenes are failing to deliver the pressure points — the architecture is right but the scenes aren't performing their function.