Lies of P companion
Use the CLI as the public entry point:
uv run --script <skill-dir>/scripts/cli.py COMMAND
- Build planning, leveling, or defaults:
fresh, then build; ask level only when needed
- Enemy matchup:
weaknesses [query]; read game_data.json when a result needs context
- Boss or area route:
bosses [query]; add --spoilers only after explicit consent
- Chapter progression:
route --chapter N; use --dlc for Overture and --spoilers explicitly
- Trophy completion:
trophies [query]; use checklist --chapter N for actionable chapter work
- Farming:
farm [stage]; avoid consumable/throwable-dependent strategies
- Meta, preferences, and community consensus:
community [query]; add --spoilers only after explicit consent
- Displayed AR or inventory candidate comparison:
compare --candidate 'NAME,PHYSICAL,ELEMENTAL,CRIT_PERCENT,WEIGHT' (repeat candidates)
- Provenance:
sources list|status|explain; audit for schema/source integrity
Default to Legendary Stalker, no Specters or summons, and spoiler-safe output. Difficulty does not alter trophy requirements. Never infer missing data: report the CLI error and request the required follow-up read. Preserve version, source, and confidence fields when presenting records. --json changes formatting only and never disables spoiler filtering.
Comparison is a deterministic displayed-AR aid, not hidden damage or DPS: displayed AR is physical + elemental; optional motion, retained lanes, and critical multiplier adjust a hit estimate. Enemy defense, hidden scaling/saturation, animation DPS, status buildup, and Fable effects are excluded. Community preferences are sentiment, not mechanics; dissent and confidence remain visible. Never claim exact hidden AR, scaling, or DPS.
Required follow-up reads
| Need |
Read |
When |
| Combat loop, weapon assembly, and stat lane |
references/combat.md |
Any build or matchup advice |
| Chapter route and weapon acquisition timing |
references/progression.md |
Any route or chapter question |
| Trophy, ending, and collectible requirements |
references/platinum.md |
Any completion or trophy question |
| Evaluation coverage and spoiler boundaries |
references/eval-coverage.md |
When validating an answer or route |
| CLI commands, arguments, and output contract |
references/cli-contract.md |
Before invoking or describing the CLI |
| Structured weapon, boss, and weakness records |
resources/game_data.json |
When CLI output needs record context |
| Source provenance and confidence |
resources/source_registry.json |
When citing or auditing a record |
| Community sentiment and calculator evidence |
resources/community.json |
When community output or comparison context needs record context |
| Trophy and chapter checklist records |
resources/platinum.json |
When trophy output needs record context |
1---2name: lies-of-p3description: Use for Lies of P 1.12.0.0 or Overture builds, routes, platinum planning, mechanics, or spoiler-safe help.4license: AGPL-3.0-or-later5---67# Lies of P companion89Use the CLI as the public entry point:10`uv run --script <skill-dir>/scripts/cli.py COMMAND`1112- Build planning, leveling, or defaults: `fresh`, then `build`; ask level only when needed13- Enemy matchup: `weaknesses [query]`; read `game_data.json` when a result needs context14- Boss or area route: `bosses [query]`; add `--spoilers` only after explicit consent15- Chapter progression: `route --chapter N`; use `--dlc` for Overture and `--spoilers` explicitly16- Trophy completion: `trophies [query]`; use `checklist --chapter N` for actionable chapter work17- Farming: `farm [stage]`; avoid consumable/throwable-dependent strategies18- Meta, preferences, and community consensus: `community [query]`; add `--spoilers` only after explicit consent19- Displayed AR or inventory candidate comparison: `compare --candidate 'NAME,PHYSICAL,ELEMENTAL,CRIT_PERCENT,WEIGHT'` (repeat candidates)20- Provenance: `sources list|status|explain`; `audit` for schema/source integrity2122Default to Legendary Stalker, no Specters or summons, and spoiler-safe output. Difficulty does not alter trophy requirements. Never infer missing data: report the CLI error and request the required follow-up read. Preserve version, source, and confidence fields when presenting records. `--json` changes formatting only and never disables spoiler filtering.2324Comparison is a deterministic displayed-AR aid, not hidden damage or DPS: displayed AR is physical + elemental; optional motion, retained lanes, and critical multiplier adjust a hit estimate. Enemy defense, hidden scaling/saturation, animation DPS, status buildup, and Fable effects are excluded. Community preferences are sentiment, not mechanics; dissent and confidence remain visible. Never claim exact hidden AR, scaling, or DPS.25## Required follow-up reads26|Need|Read|When|27|---|---|---|28|Combat loop, weapon assembly, and stat lane|`references/combat.md`|Any build or matchup advice|29|Chapter route and weapon acquisition timing|`references/progression.md`|Any route or chapter question|30|Trophy, ending, and collectible requirements|`references/platinum.md`|Any completion or trophy question|31|Evaluation coverage and spoiler boundaries|`references/eval-coverage.md`|When validating an answer or route|32|CLI commands, arguments, and output contract|`references/cli-contract.md`|Before invoking or describing the CLI|33|Structured weapon, boss, and weakness records|`resources/game_data.json`|When CLI output needs record context|34|Source provenance and confidence|`resources/source_registry.json`|When citing or auditing a record|35|Community sentiment and calculator evidence|`resources/community.json`|When `community` output or comparison context needs record context|36|Trophy and chapter checklist records|`resources/platinum.json`|When trophy output needs record context|
anntnzrb