self-modifying-code

Tiers, checkpoints, and off-limits paths for an agent editing the app's own source. Use when designing UI for agent editability, deciding whether a file is safe for the agent to change, or when tempted to patch config, `.env`, or an `@agent-native/*` package. Do not load it for ordinary source edits.

Quality

86%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

88%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is concrete, executable, and well-sequenced, with a strong validation-gated checkpoint loop for risky edits. Structure is clean; the main weakness is mild redundancy between the taxonomy, upgrade guidance, and 'Don't' list.

Suggestions

Consolidate the 'Don't' list with the Tier 4 taxonomy so each off-limits rule appears once, reducing redundancy and token cost.

Move the lengthy `@agent-native/*` patch/override enumeration into a short reference file or trim to the highest-signal examples now that no bundle files offload detail.

Cross-link 'Related Skills' entries inline at first mention (e.g. upgrade-agent-native) so navigation is one-level and signposted where the reader needs it.

Dimension	Reasoning	Score
Conciseness	Mostly lean and assumes Claude's competence (no generic explanations of git or config), but the 'Don't' list reiterates points already made in the taxonomy and upgrade paragraphs, adding mild redundancy.	4 / 5
Actionability	Provides executable TypeScript snippets, exact commands (`pnpm typecheck && pnpm lint`, `git checkout -- <file>`, `npx @agent-native/core@latest upgrade`), and a concrete tier table with clear per-tier actions.	5 / 5
Workflow Clarity	The Git Checkpoint Pattern is a fully sequenced five-step loop with an explicit validation checkpoint (`pnpm typecheck && pnpm lint`) and a revert-on-failure feedback path, satisfying the destructive-edit feedback-loop requirement.	5 / 5
Progressive Disclosure	Well-organized sections with clear navigation to peer skills via the 'Related Skills' list, but the off-limits detail and 'Don't' section are somewhat long for the body when no bundle files exist to offload them.	4 / 5
	Total	18 / 20 Passed

Description

85%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, targeted description that pairs a clear purpose with explicit 'Use when' triggers and a 'do not load' negative trigger. It is slightly jargon-heavy and the concrete-action list is more declarative than enumerative.

Dimension	Reasoning	Score
Specificity	Names concrete actions like 'designing UI for agent editability' and 'deciding whether a file is safe for the agent to change', but the action vocabulary is modest rather than a comprehensive multi-action list.	4 / 5
Completeness	Explicitly answers 'what' (tiers/checkpoints/off-limits paths for an agent editing source) and 'when' ('Use when designing UI...', 'Do not load it for ordinary source edits') with concrete trigger phrases.	5 / 5
Trigger Term Quality	Includes natural trigger phrases ('designing UI for agent editability', 'tempted to patch config, `.env`, or an `@agent-native/*` package') plus file extensions/paths, though it leans on project-specific jargon over broadly natural user terms.	4 / 5
Distinctiveness Conflict Risk	The `@agent-native/*` and `.env` patching niche is fairly distinct, but it overlaps conceptually with the referenced customizing/upgrade skills for closely related edit decisions.	4 / 5
	Total	17 / 20 Passed

Validation

81%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning
metadata_field	'metadata' should map string keys to string values	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	13 / 16 Passed

Repository: BuilderIO/agent-native
Path: .agents/skills/self-modifying-code/SKILL.md
Commit: bb282b1

Reviewed: 5 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.