mutate
Skill6 starsChanged 14 days agoApache-2.0View on GitHub
Change persistent state with checkpoint and rollback support. Use when modifying files, updating databases, changing configuration, or any operation that permanently alters state.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- OHOKAYYEAH/claude-skillstheir own version, first seen Oct 13, 2026
- Claritune/mutantxtheir own version, first seen Aug 2, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- Claritune/mutantx★ 2same name
- OHOKAYYEAH/claude-skills★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository42 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- attributeSkill
- auditSkill
- gap-analysis-workflowSkill
- checkpointSkill
- classifySkill
- compareSkill
- constrainSkill
- critiqueSkill
- debug-workflowSkill
- decomposeSkill
- delegateSkill
- detectSkill
- digital-twin-bootstrapSkill
- digital-twin-sync-workflowSkill
- discover-capabilitiesSkill
- discoverSkill
- executeSkill
- explainSkill
- generateSkill
- groundSkill
- inquireSkill
- integrateSkill
- invokeSkill
- measureSkill
- observeSkill
- persistSkill
- perspective-validationSkill
- planSkill
- predictSkill
- recallSkill
- receiveSkill
- retrieveSkill
- rollbackSkill
- searchSkill
- sendSkill
- simulateSkill
- stateSkill
- synchronizeSkill
- transformSkill
- transitionSkill
- verifySkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
16 rules. About 0 are common, 16 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈16
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 40
Verify checkpoint exists: Ensure recovery is possible
- Distinctive: line 42
Verify checkpoint covers the target
- Distinctive: line 43
STOP if no valid checkpoint
- Distinctive: line 46
Read current state of target
- Distinctive: line 48
Document what specifically will change
- Distinctive: line 52
Use appropriate tool (Edit for files, Bash for commands)
- Distinctive: line 60
Record mutation: Document what changed
- Distinctive: line 223
Verification tools: Read (to verify state changes)
- Distinctive: line 233
NEVER mutate without valid checkpoint
- Distinctive: line 234
Always capture previous state before modification
- Distinctive: line 235
Verify change matches intended operation
- Distinctive: line 237
Stop immediately on unexpected errors
- Distinctive: line 238
Do not chain multiple mutations without intermediate checkpoints
- Distinctive: line 253
NEVER mutate without checkpoint
- Distinctive: line 254
Never combine multiple risky mutations
- Distinctive: line 255
Avoid mutation in exploratory workflows
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Intent
Make controlled changes to persistent state with safety guarantees. This is the primary capability for modifications and REQUIRES a checkpoint for rollback capability.
CRITICAL: This capability requires checkpoint to have been called first.
Success criteria:
- State change applied correctly
- Previous state captured for rollback
- Change is verifiable
- Audit trail recorded
Compatible schemas:
schemas/output_schema.yaml
Inputs
| Parameter | Required | Type | Description |
target | Yes | string | What to modify (file path, resource identifier) |
operation | Yes | object | The modification to apply |
checkpoint_id | No | string | Checkpoint ID for rollback (from checkpoint capability) |
Inputs not listed here are outside the capability's contract: a workflow step that binds them gets W604 (STANDARD v1.1 §5.5).
Procedure
- Verify checkpoint exists: Ensure recovery is possible
- Confirm checkpoint_id references valid checkpoint
- Verify checkpoint covers the target
- STOP if no valid checkpoint
- Capture previous state: Record what will change
- Read current state of target
- Store for comparison and rollback
- Document what specifically will change
- Apply mutation: Make the change
- Execute the specified operation
- Use appropriate tool (Edit for files, Bash for commands)
- Capture any errors during modification
- Verify change: Confirm mutation succeeded
- Read new state
- Compare to expected result
- Note any unexpected side effects
- Record mutation: Document what changed
- Store diff between previous and new state
- Record timestamp and operation details
- Link to checkpoint for audit trail
- Return result: Provide complete mutation record
- Include success/failure status
- Provide before/after state
- Reference checkpoint for rollback
Output Contract
Return a structured object:
success: boolean # Whether mutation completed
target: string # What was modified
operation: string # What was done
previous_state:
snapshot: any # State before mutation
captured_at: string # When captured
new_state:
snapshot: any # State after mutation
verified_at: string # When verified
diff:
type: string # unified, structural, summary
content: string # The actual diff
checkpoint_id: string # For rollback reference
rollback_command: string # How to undo this change
evidence_anchors: ["file:line", "command:output"]Field Definitions
| Field | Type | Description |
success | boolean | Whether mutation completed successfully |
target | string | Resource that was modified |
operation | string | Description of the change |
previous_state | object | State before modification |
new_state | object | State after modification |
diff | object | Difference between states |
checkpoint_id | string | Reference for rollback |
rollback_command | string | How to undo the change |
evidence_anchors | array | References to the changed location and the command output |
Examples
Example 1: File Modification
Input:
target: "src/services/user_service.rb"
operation:
type: "edit"
description: "Add null check for user parameter"
changes:
- line: 42
old: "user.email"
new: "user&.email"
checkpoint_id: "chk-20240115-001"Output:
success: true
target: "src/services/user_service.rb"
operation: "Added safe navigation operator for null user handling"
previous_state:
snapshot: "def get_email(user)\n user.email\nend"
captured_at: "2024-01-15T10:30:00Z"
new_state:
snapshot: "def get_email(user)\n user&.email\nend"
verified_at: "2024-01-15T10:30:01Z"
diff:
type: "unified"
content: |
@@ -42 +42 @@
- user.email
+ user&.email
checkpoint_id: "chk-20240115-001"
rollback_command: "git checkout src/services/user_service.rb"
evidence_anchors:
- "src/services/user_service.rb:42"Example 2: Database Mutation
Input:
target: "users table"
operation:
type: "sql"
description: "Add status column with default value"
sql: "ALTER TABLE users ADD COLUMN status VARCHAR(20) DEFAULT 'active'"
checkpoint_id: "chk-20240115-002"Output:
success: true
target: "users table"
operation: "Added status column with default 'active'"
previous_state:
snapshot:
columns: ["id", "email", "created_at"]
captured_at: "2024-01-15T11:00:00Z"
new_state:
snapshot:
columns: ["id", "email", "created_at", "status"]
verified_at: "2024-01-15T11:00:05Z"
diff:
type: "structural"
content: "+ status VARCHAR(20) DEFAULT 'active'"
checkpoint_id: "chk-20240115-002"
rollback_command: "ALTER TABLE users DROP COLUMN status"
evidence_anchors:
- "command:psql:schema_after"Example 3: Mutation Failure
Input:
target: "config/database.yml"
operation:
type: "edit"
description: "Update connection pool size"
checkpoint_id: "chk-20240115-003"Output:
success: false
target: "config/database.yml"
operation: "Update connection pool size (FAILED)"
previous_state:
snapshot: "pool: 5"
captured_at: "2024-01-15T12:00:00Z"
new_state:
snapshot: null
verified_at: null
diff:
type: "none"
content: "Mutation failed: file locked by another process"
checkpoint_id: "chk-20240115-003"
rollback_command: "N/A - mutation did not complete"
evidence_anchors:
- "error:file_locked"
error:
message: "File locked by another process"
recoverable: true
suggestion: "Retry after process releases lock"Verification
- Checkpoint exists and is valid
- Previous state captured before mutation
- New state verified after mutation
- Diff accurately reflects change
- Rollback command is correct
Verification tools: Read (to verify state changes)
Safety Constraints
mutation: truerequires_checkpoint: truerequires_approval: truerisk: high
Capability-specific rules:
- NEVER mutate without valid checkpoint
- Always capture previous state before modification
- Verify change matches intended operation
- Provide working rollback mechanism
- Stop immediately on unexpected errors
- Do not chain multiple mutations without intermediate checkpoints
Composition Patterns
Commonly follows:
checkpoint- REQUIRED before any mutationplan- Mutations execute planned changesverify- Verify preconditions before mutating
Commonly precedes:
verify- Verify mutation resultsaudit- Record mutation for audit trailrollback- Undo mutation if verification fails
Anti-patterns:
- NEVER mutate without checkpoint
- Never combine multiple risky mutations
- Avoid mutation in exploratory workflows
Workflow references:
- See
reference/workflow_catalog.yaml#debug_code_changefor mutation in bug fixes - See
reference/workflow_catalog.yaml#digital_twin_sync_loopfor mutation in sync loops
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

