fix(icrc-ledger): fix deploy config, add pitfalls, and evaluations

[marc0olo]

Summary

Closes #106

• Fix broken deploy config: Replace invalid type: custom recipe with build.steps + type: pre-built and init_args: { path, format } syntax (verified with actual deployment)
• Fix minting account pitfalls: Separate minting/funded principals in init args, document icrc2_approve trap (#9) and burn/mint zero-fee semantics (#10)
• Rewrite fee guidance: Per-ledger fees, icrc1_fee lookup, BadFee handling (pitfall #1)
• Add evaluations: 4 output evals (Motoko transfer, Rust ICRC-2, burn fee semantics, deploy config) + 16 trigger evals — all verified with eval runs showing clear skill uplift
• Clean up: Remove low-value verification sections, rename to "Common Pitfalls", bump core to 2.3.1, add icrc1_metadata/icrc1_supported_standards to actor interface, show multiple initial_balances entries

Eval results

Eval	With skill	Without	Delta
ICRC-1 transfer in Motoko	5/5	2/5	+3
ICRC-2 approve+transferFrom in Rust	5/5	3/5	+2
Burn fee semantics	3/3	3/3	0
Deploy an ICRC-1 ledger	5/5	3/5	+2
Triggers	16/16	—	perfect

Not in scope

The frontend ICRC-2 approval flow (inline IDL factories, calling ledgers from JavaScript) is deferred — it's a generic canister interaction pattern that belongs in a future canister-calls skill. The token custody pattern (approve flow vs subaccount deposits) was evaluated but dropped since agents already know it without the skill (3/3 baseline).

[github-actions]

Skill Validation Report

Validating skill: /home/runner/work/icskills/icskills/skills/icrc-ledger

Structure

• Pass: SKILL.md found

Frontmatter

• Pass: name: "icrc-ledger" (valid)
• Pass: description: (384 chars)
• Warning: description has 8 comma-separated segments, most very short — this looks like a keyword list; per the spec, the description should concisely describe what the skill does and when to use it
• Pass: license: "Apache-2.0"
• Pass: compatibility: (16 chars)
• Pass: metadata: (2 entries)

Markdown

• Pass: no unclosed code fences found

Tokens

File	Tokens
SKILL.md body	4,365
Total	4,365

Content Analysis

Metric	Value
Word count	2,145
Code block ratio	0.61
Imperative ratio	0.11
Information density	0.36
Instruction specificity	1.00
Sections	12
List items	17
Code blocks	7

Contamination Analysis

Metric	Value
Contamination level	low
Contamination score	0.20
Primary language category	config
Scope breadth	4

• Warning: Language mismatch: shell, systems (2 categories differ from primary)

Result: 1 warning

Project Checks

✓ Project checks passed for 1 skills (0 warnings)