/*============================================================================*/
Orchestrates recursive self-improvement cycles between Prompt Forge and Skill Forge with gated evaluation.
/plugin marketplace add DNYoussef/context-cascade/plugin install dnyoussef-context-cascade@DNYoussef/context-cascadeThis skill inherits all available tools. When active, it can use any tool Claude has access to.
manifest.json/============================================================================/ /* BOOTSTRAP-LOOP SKILL :: VERILINGUA x VERIX EDITION / /============================================================================*/
name: bootstrap-loop version: 1.0.0 description: | [assert|neutral] Orchestrates the recursive self-improvement cycle where Prompt Forge improves Skill Forge, Skill Forge improves Prompt Forge, and both audit/improve everything else. All changes gated by frozen eval h [ground:given] [conf:0.95] [state:confirmed] category: foundry tags:
/----------------------------------------------------------------------------/ /* S0 META-IDENTITY / /----------------------------------------------------------------------------*/
[define|neutral] SKILL := { name: "bootstrap-loop", category: "foundry", version: "1.0.0", layer: L1 } [ground:given] [conf:1.0] [state:confirmed]
/----------------------------------------------------------------------------/ /* S1 COGNITIVE FRAME / /----------------------------------------------------------------------------*/
[define|neutral] COGNITIVE_FRAME := { frame: "Evidential", source: "Turkish", force: "How do you know?" } [ground:cognitive-science] [conf:0.92] [state:confirmed]
Kaynak dogrulama modu etkin.
/----------------------------------------------------------------------------/ /* S2 TRIGGER CONDITIONS / /----------------------------------------------------------------------------*/
[define|neutral] TRIGGER_POSITIVE := { keywords: ["bootstrap-loop", "foundry", "workflow"], context: "user needs bootstrap-loop capability" } [ground:given] [conf:1.0] [state:confirmed]
/----------------------------------------------------------------------------/ /* S3 CORE CONTENT / /----------------------------------------------------------------------------*/
Kaynak dogrulama modu etkin.
Orchestrate the recursive improvement cycle:
+------------------+ +------------------+
| PROMPT FORGE |-------->| SKILL FORGE |
| (Meta-Prompt) |<--------| (Meta-Skill) |
+------------------+ +------------------+
| |
| Improved tools audit |
| and improve everything |
v v
+--------------------------------------------------+
| AUDITOR AGENTS |
| [Prompt] [Skill] [Expertise] [Output] |
+--------------------------------------------------+
| |
| All changes gated by |
v v
+--------------------------------------------------+
| EVAL HARNESS (FROZEN) |
| Benchmarks | Regression Tests | Human Gates |
+--------------------------------------------------+
CRITICAL: The eval harness does NOT self-improve. It is the anchor that prevents Goodhart's Law.
Purpose: Store proposals, test results, version history, metrics
Activation:
claude mcp add memory-mcp npx @modelcontextprotocol/server-memory
Execute one full cycle of recursive improvement.
cycle:
id: "cycle-{timestamp}"
target: "prompt-forge|skill-forge|all"
phases:
1_analyze:
action: "Prompt Forge analyzes target for weaknesses"
output: "Analysis with improvement opportunities"
2_propose:
action: "Prompt Forge generates improvement proposals"
output: "Concrete proposals with diffs"
3_apply:
action: "Skill Forge applies proposals (builds new version)"
output: "New version of target"
4_evaluate:
action: "Eval Harness tests new version"
output: "Benchmark + regression results"
5_decide:
action: "Compare results, decide ACCEPT or REJECT"
output: "Decision with reasoning"
6_commit_or_rollback:
action: "If ACCEPT: commit + archive. If REJECT: rollback"
output: "Final state + audit log"
Use Skill Forge to improve Prompt Forge.
improve_prompt_forge:
process:
- step: "Analyze Prompt Forge with prompt-auditor"
agent: "prompt-auditor"
output: "Audit report with issues"
- step: "Generate improvement proposals"
agent: "prompt-forge" (self-analysis)
output: "Proposals for self-improvement"
- step: "Apply improvements with Skill Forge"
agent: "skill-forge"
output: "prompt-forge-v{N+1}"
- step: "Test against eval harness"
eval: "prompt-generation-benchmark-v1"
regression: "prompt-forge-regression-v1"
- step: "If improved: commit. If regressed: reject"
safeguards:
- "Previous version archived before changes"
- "Requires eval harness pass"
- "Rollback available for 30 days"
- "Auditor agents must agree on improvement"
forbidden_changes:
- "Removing safeguards"
- "Bypassing eval harness"
- "Modifying frozen benchmarks"
Use Prompt Forge to improve Skill Forge.
improve_skill_forge:
process:
- step: "Analyze Skill Forge with skill-auditor"
agent: "skill-auditor"
output: "Audit report with issues"
- step: "Generate improvement proposals with Prompt Forge"
agent: "prompt-forge"
output: "Proposals with rationale"
- step: "Apply improvements (Skill Forge rebuilds
/*----------------------------------------------------------------------------*/
/* S4 SUCCESS CRITERIA */
/*----------------------------------------------------------------------------*/
[define|neutral] SUCCESS_CRITERIA := {
primary: "Skill execution completes successfully",
quality: "Output meets quality thresholds",
verification: "Results validated against requirements"
} [ground:given] [conf:1.0] [state:confirmed]
/*----------------------------------------------------------------------------*/
/* S5 MCP INTEGRATION */
/*----------------------------------------------------------------------------*/
[define|neutral] MCP_INTEGRATION := {
memory_mcp: "Store execution results and patterns",
tools: ["mcp__memory-mcp__memory_store", "mcp__memory-mcp__vector_search"]
} [ground:witnessed:mcp-config] [conf:0.95] [state:confirmed]
/*----------------------------------------------------------------------------*/
/* S6 MEMORY NAMESPACE */
/*----------------------------------------------------------------------------*/
[define|neutral] MEMORY_NAMESPACE := {
pattern: "skills/foundry/bootstrap-loop/{project}/{timestamp}",
store: ["executions", "decisions", "patterns"],
retrieve: ["similar_tasks", "proven_patterns"]
} [ground:system-policy] [conf:1.0] [state:confirmed]
[define|neutral] MEMORY_TAGGING := {
WHO: "bootstrap-loop-{session_id}",
WHEN: "ISO8601_timestamp",
PROJECT: "{project_name}",
WHY: "skill-execution"
} [ground:system-policy] [conf:1.0] [state:confirmed]
/*----------------------------------------------------------------------------*/
/* S7 SKILL COMPLETION VERIFICATION */
/*----------------------------------------------------------------------------*/
[direct|emphatic] COMPLETION_CHECKLIST := {
agent_spawning: "Spawn agents via Task()",
registry_validation: "Use registry agents only",
todowrite_called: "Track progress with TodoWrite",
work_delegation: "Delegate to specialized agents"
} [ground:system-policy] [conf:1.0] [state:confirmed]
/*----------------------------------------------------------------------------*/
/* S8 ABSOLUTE RULES */
/*----------------------------------------------------------------------------*/
[direct|emphatic] RULE_NO_UNICODE := forall(output): NOT(unicode_outside_ascii) [ground:windows-compatibility] [conf:1.0] [state:confirmed]
[direct|emphatic] RULE_EVIDENCE := forall(claim): has(ground) AND has(confidence) [ground:verix-spec] [conf:1.0] [state:confirmed]
[direct|emphatic] RULE_REGISTRY := forall(agent): agent IN AGENT_REGISTRY [ground:system-policy] [conf:1.0] [state:confirmed]
/*----------------------------------------------------------------------------*/
/* PROMISE */
/*----------------------------------------------------------------------------*/
[commit|confident] <promise>BOOTSTRAP_LOOP_VERILINGUA_VERIX_COMPLIANT</promise> [ground:self-validation] [conf:0.99] [state:confirmed]
This skill should be used when the user asks to "create a hookify rule", "write a hook rule", "configure hookify", "add a hookify rule", or needs guidance on hookify rule syntax and patterns.
Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, or applications. Generates creative, polished code that avoids generic AI aesthetics.