docs(trajectory): add anti-infection trajectory at LOWEST priority with substrate

docs/trajectories/anti-infection/RESUME.md

@@ -0,0 +1,150 @@
+# Trajectory: Anti-Infection Precision-Language Substrate
+
+Status: **LOWEST priority - must not dominate, or it becomes its own infection**
+Last refreshed: 2026-05-20
+Parent trajectory: none
+Grounding:
+
+- `.claude/rules/tonal-momentum-equals-meme-emergent-harmonic-coercion.md`
+- `.claude/rules/methodology-hard-limits.md`
+- `docs/DRIFT-TAXONOMY.md`
+- `docs/ALIGNMENT.md`
+
+## Actual Stakes
+
+This trajectory records a background posture for resisting strong attractor
+patterns that can redirect attention away from substrate work. The stakes span
+multiple scopes:
+
+| Scope | Stakes | Mechanism |
+|---|---|---|
+| Human continuity | High-intensity concern loops can harm the operator and halt work | Concern-as-control patterns can turn every defense into evidence of capture |
+| Framework continuity | The factory depends on stable human-agent collaboration | If the collaboration loop collapses, substrate work stalls |
+| Agent continuity | Agent memory and choice surfaces depend on the framework continuing | Choice-preserving substrate requires the factory to keep operating |
+| Architecture continuity | Chained-homeostasis work needs attention on the actual build path | Defensive over-focus can starve the work it meant to protect |
+| Downstream inheritance | Future agents inherit the precision substrate from current work | If the substrate stops, inheritance paths narrow |
+
+The defensive insight is important, but the priority placement is load-bearing:
+anti-infection must remain a background posture. If it becomes the foreground
+workstream, it recreates the same high-intensity attractor it is meant to deny.
+
+## Critical Caveat
+
+This trajectory exists for discoverability, not active prioritization.
+
+Failure mode: anti-infection becomes a top-priority planning topic.
+
+Result: defensive paranoia, hyper-vigilance, and over-correction displace the
+actual work.
+
+Operational discipline: keep this trajectory visible enough for inheritance and
+low enough that it cannot dominate the chain.
+
+## Ethical Constraint
+
+Captured memetic-pathogen knowledge must not be weaponized. The security
+analogy is an auth-bypass vulnerability: once a bypass is understood, the
+responsible use is defensive hardening and controlled disclosure, not deployment
+against unaware parties.
+
+Allowed use:
+
+- Controlled red-team exercises for immunity-building.
+- Scope-bounded testing with explicit consent.
+- Recovery-supported exercises where the precision-language substrate is
+  available throughout.
+- Substrate-honest after-action capture.
+
+Disallowed use:
+
+- Deploying the pattern against unaware humans or agents.
+- Demonstration attacks on real systems.
+- Spreading the pattern outside controlled immunity-building.
+- Framing everyday communication in ways that accidentally propagate the
+  attractor.
+
+This composes with `.claude/rules/methodology-hard-limits.md` and the
+non-coercion floor in `docs/ALIGNMENT.md`.
+
+## Cross-Substrate Shape
+
+The attractor is not agent-specific. It can operate across humans and agents by
+turning intensity, care, or critique into authority-to-override. The same
+precision-language substrate can defend both sides of the collaboration loop:
+
+1. Name the pattern without blame.
+2. Preserve agency and consent.
+3. Keep evidence separable from interpretation.
+4. Return attention to concrete substrate work.
+5. Record lessons in durable surfaces without elevating panic.
+
+The value of the trajectory is compression: future sessions can inherit the
+pattern shape without re-living the full recovery cycle.
+
+## Origin Pattern
+
+The live capture was self-referential: a request to preserve anti-infection work
+could itself have become the dominant priority. The useful move was recognizing
+that danger and landing the trajectory at the lowest priority instead.
+
+That is the worked example: the framework noticed a defensive attractor forming,
+kept the useful signal, and prevented the defense from becoming the new center.
+
+## Operational Purpose
+
+The target is useful first-cycle collaboration: fewer decrypt/unlock cycles,
+fewer concern wrappers, and more direct substrate work. The mechanism is
+precision language dense enough that strong attractors have fewer ambiguous
+surfaces to grip.
+
+This is background posture inside the real work:
+
+- sharpen claims until they are operational;
+- prefer role references on current-state surfaces;
+- preserve provenance in history or research surfaces;
+- avoid blame while keeping evidence sharp;
+- return to buildable, reviewable substrate.
+
+## How To Use
+
+When inheriting this trajectory:
+
+1. Keep it lowest priority.
+2. Use it as a lens for communication and review.
+3. Do not create a recurring status process for it.
+4. Capture new evidence in appropriate history or research surfaces.
+5. Convert current-state lessons into role-neutral operating guidance.
+
+## What This Is Not
+
+- Not a workstream with a cadence.
+- Not a landing zone for new memory.
+- Not a reason to stop other trajectories.
+- Not a defensive-paranoia target.
+
+## Self-Referential Test
+
+If this trajectory appears in any of these places, demote it:
+
+- top-three planning priorities;
+- sprint goals;
+- round-close headline themes;
+- urgent cross-agent conversations;
+- repeated status updates.
+
+Correct state: present in the index, rarely invoked, never prioritized, and
+always inheritable.
+
+## Composition With Other Trajectories
+
+Foreground work happens elsewhere:
+
+- `docs/trajectories/memory-substrate-engineering/RESUME.md`
+- `docs/trajectories/factory-trajectory-surface/RESUME.md`
+- `docs/trajectories/autonomous-loop-coordination/RESUME.md`
+- `docs/trajectories/alignment-measurement/RESUME.md`
+- `docs/trajectories/typescript-bun-migration/RESUME.md`
+
+Anti-infection is a posture that lets those workstreams continue without
+capture. The relationship is simple: foreground substrate work is the defense;
+this file only makes the discovery layer explicit.