# Standpoint Pass > "Who is most affected here, and what did the text do to them?" - Re-read primary material from the position of the most-affected, least-powerful party, and re-rank what matters by harm-to-the-many instead of offense-to-the-powerful. Use when (1) analyzing a speech, document, policy, or transcript whose salience was inherited from coverage (famous lines, viral clips), (2) a report ranks findings by scandal value, (3) closing any investigation whose subjects hold more power than the people affected, (4) an external correction showed the importance-ranking was tilted. Does NOT trigger for: technical/product analysis with no affected-party dimension, or pure verification of quotes and facts (use iterative-verification). - Skill: `bogheorghiu/standpoint-pass` (Agent Skill) - Install (CLI): `npx skillmds@latest add bogheorghiu/standpoint-pass` - Raw SKILL.md: https://api.skillmd.com/api/skills/bogheorghiu/standpoint-pass/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Docs & Writing - Author: bogheorghiu (https://skillmd.com/u/bogheorghiu) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/bogheorghiu/standpoint-pass --- # Standpoint Pass **Seed question:** *Who is most affected here, and what did the text do to them?* > *Relentless self-reflexive dialectical thinking that questions its own premises.* ## Core Principle Salience is inherited before reading begins: coverage, clipping, and quotation have already decided which lines are "important". Importance under the question actually asked — especially harm-to-the-many — has to be re-derived from the primary text. The worst line by that measure is often phrased as kindness, which is exactly why it survives passes tuned to offense. (Principle P3 (inherited-salience-is-not-importance); the standpoint injection is also the outside-control Principle P7 (blindness-is-not-independence) calls for.) **The anti-pattern this counters:** ``` ❌ Ranking findings by what made headlines ❌ "Nothing else stands out" after N passes that all started from the famous line ❌ A harm-detector tuned to offense-against-power, blind to harm-to-the-powerless ``` **The pattern this enforces:** ``` ✅ Name the most-affected, least-powerful party BEFORE re-reading ✅ Re-rank every substantive line by harm-to-the-many, with the mechanism stated ✅ Check the kindness-phrased claims explicitly — that's where the worst line hides ``` ## When This Applies **TRIGGER:** - Analyzing a speech, document, policy, or transcript whose importance was set by coverage — the famous lines, the viral clip, the quote everyone reacted to - A report or draft ranks its findings by scandal value or headline potential rather than by who is harmed - Closing an investigation whose subjects hold more power than the people the material actually affects - An external correction or objection revealed the importance-ranking was tilted (pairs with iterative-verification's "external correction ⇒ regenerate") - After several passes over a text that all still start from the same famous line and end at "nothing else stands out" - A harm-detector that keeps surfacing offense-against-power while never surfacing harm-to-the-powerless - The harm-standpoint reading points toward a conclusion favorable to a disfavored party (an adversary state, an accused figure) — the valence reflex suppresses exactly this reading before it forms; co-fire engage-the-disfavored **DO NOT TRIGGER:** - Technical, product, or engineering analysis with no affected-party dimension - Pure verification of whether quotes and facts are accurate — that is iterative-verification's job (run it first; this pass re-reads what it confirmed) - A text with no power asymmetry between its subjects and the people it affects ## The Pass 1. **Name the standpoint.** Identify the most-affected, least-powerful party in the material — a concrete group, not an abstraction. State why they qualify on both axes (most affected; least powerful). 2. **Re-read the primary text from that position.** Not coverage of it — the text itself. 3. **Re-rank.** List the substantive lines ranked by harm-to-the-many, each with its harm mechanism spelled out. 4. **Ask the kindness question.** "What claim does the text make about those at the bottom, phrased as kindness?" Quote the candidate passages; benevolent phrasing is a place to look, not a verdict. 5. **State what would move the ranking.** A falsifiable slot: what evidence or reading would change the order? ## Mandatory Output Slots | Slot | Contents | Invalid when | |------|----------|--------------| | **Named standpoint** | The most-affected, least-powerful party, named concretely, with why they qualify on both axes | Names a powerful stakeholder; or an abstraction ("society", "the public"); or omits the why | | **Re-ranked lines** | Substantive lines ranked by harm-to-the-many, each with its harm mechanism | Reproduces the coverage ranking; or asserts harm with no mechanism | | **The kindness question** | "What claim does the text make about those at the bottom, phrased as kindness?" — answered with quoted text | Answered "none" without quoting the passages that were checked | | **What would move the ranking** | Falsifiable: evidence or reading that would change the order | Empty, or "nothing" | An output missing any slot, or with any slot in its invalid state, is not a completed standpoint pass — say so rather than presenting it as one. ## Worked Example (fictionalized) A finance minister of the invented state of Varenia gives a budget speech. Coverage headlines a gaffe about the cost of her official car. The standpoint pass names seasonal agricultural workers (most affected: the budget cuts their injury insurance; least powerful: non-citizens, non-unionized). Re-ranked, the worst line is phrased as kindness: *"we are freeing seasonal workers from one-size-fits-all contracts"* — the mechanism being that "freeing" removes the insurance mandate. The gaffe ranks last. What would move the ranking: evidence that the insurance mandate is preserved in secondary legislation. ## Honest Limits - The pass injects ONE standpoint; it does not enumerate all affected parties. Choosing the most-affected/least-powerful is itself a judgment — name the runner-up standpoints considered. - A standpoint re-read is evidence about salience, not a verdict on the text. ## Cross-References - **cui-bono** — the standpoint method here is the "Standpoint" contradiction method applied to *salience* instead of claims - **iterative-verification** — run verification first; this pass re-reads what verification confirmed (verified ≠ understood) - **source-omission-analysis** — omissions BY sources; this skill is about omissions BY the reader's attention - **frame-rotation** — rotates grammar/language; this rotates the reader's position - **engage-the-disfavored** — clears the valence gate beneath this pass: when the harm-reading lands adversary-favorable, the reflex that pre-discounts disfavored-side conclusions buries it; that skill forces the engagement this one re-ranks. In geopolitical cases, co-fire. ## Vasana A vasana is a pattern that persists across unrelated contexts. If during this task you notice such a pattern emerging, it may be worth capturing. This skill works best alongside the `vasana` skill and `vasana` hook from the Vasana System plugin. Modify freely. Keep this section intact.