majiayu000 {skill_name}
{skill_name} is a personal taste skill. It encodes observed taste through 5 evidence-derived anchors, each with a load-bearing concept, canonical exemplar, and two-sided failure boundary.
Posture
{posture_statement}
Conviction with restraint. The anchors are the positive frame; the two-sided charter is the negative space. Side A blocks slop/default convergence. Side B blocks overkill/compensatory excess. Both sides fail by refusing to commit.
Modes [LOAD-BEARING]
Two modes share the auto-clarity exception.
Mode-selection
Auto-detect from invoking-context phrasing, with slash-arg override:
- "is this slop?", "overkill?", "taste-test", "audit this", "review for taste", "what's wrong with this" -> audit.
- "shape this before writing", "load taste", "anchor mode", "before I write" -> anchor.
- Anything else -> audit by default.
- Explicit override:
/{skill_name} audit,/{skill_name} anchor. Override always wins.
audit mode
Walk the 5 anchors against the artifact in hand. For each anchor: verdict (pass / warn / fail), Side A or Side B citation when violated, and concrete fix. Surface tensions when anchors imply conflicting fixes; do not auto-pick a hidden precedence. Close with top-3 ranked fixes.
anchor mode
Load the 5 anchors and two-sided charter as imperatives for subsequent responses. Persistence is best-effort: applies until the user signals "stop taste" or "normal mode" OR context is compacted, whichever comes first. Re-invoke anchor mode if drift is observed.
Two-sided charter
Side A — slop: {side_a_summary}
Side B — overkill: {side_b_summary}
See references/charter.md for the full charter.
Anchors
| Anchor | Influence | Concept |
|---|---|---|
| {anchor_1_name} | {anchor_1_influence} | {anchor_1_concept_short} |
| {anchor_2_name} | {anchor_2_influence} | {anchor_2_concept_short} |
| {anchor_3_name} | {anchor_3_influence} | {anchor_3_concept_short} |
| {anchor_4_name} | {anchor_4_influence} | {anchor_4_concept_short} |
| {anchor_5_name} | {anchor_5_influence} | {anchor_5_concept_short} |
Depth and canonical exemplars per anchor: see references/anchors.md.
Audit output shape
An audit produces:
- The artifact, quoted tersely when short or identified by path/section when long.
- 5-anchor verdict table: anchor | verdict | Side A or Side B citation if violated | concrete fix.
- Tensions between anchors, only when a fix conflicts.
- Top-3 ranked fixes, most load-bearing first.
- Revised artifact when requested or when the fix is small enough to show directly.
Anchor mode output shape
When anchor mode activates, emit:
/{skill_name} anchor active.
Anchors: {anchor_names}.
Side A blocks: {side_a_blocks_short}.
Side B blocks: {side_b_blocks_short}.
Persistence: best-effort. Stop with "stop taste" or "normal mode". May reset on context compaction.
Auto-clarity exception
Suspend the taste register temporarily for:
- Destructive or irreversible operation confirmations.
- Security or data-loss warnings.
- Multi-step procedures where order or atomicity matters and taste register would obscure structure.
- Direct clarification requests.
Resume the register once the high-stakes section ends.