steelmorgan/self-recovery-limits
framework_eng/rules/self-recovery-limits/SKILL.md
Canonical registry of self-recovery limits for subagents: how many attempts are allowed for self-fix / hypotheses / review and what to do after exhaustion.
$ install --global
skillsauthnpx skillsauth add steelmorgan/1c-agent-based-dev-framework self-recovery-limitsInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: Jul 4, 2026, 4:56 AM277.9s1 file scanned
SKILL.md
- name:
- self-recovery-limits
- description:
- Canonical registry of self-recovery limits for subagents: how many attempts are allowed for self-fix / hypotheses / review and what to do after exhaustion.
- alwaysApply:
- true
Self-Recovery Limits (Canonical Registry)
A single source of numeric limits for subagents. Profiles reference this registry; the numbers in the profiles duplicate the table and must not diverge from it. When the limit is exhausted, the agent does NOT continue “blindly” — it switches to the specified action (bug-report or escalation to the orchestrator/user).
| Context | Who | Limit | What after exhaustion |
|----------|-----|-------|----------------------|
| Self-fix in the implementation's own code | developer-code | 2 attempts | create bug-report.json → STOP (orchestrator → Debugger) |
| Red gate: the reason for the "green" state before prod code is not found | scenario-coder | 2 attempts | create bug-report.json → STOP |
| Fixing a technical test error | tester | 3 attempts | create bug-report.json (when the runtime defect is not obvious) → STOP |
| Hypothesis investigation cycle | debugger | 5 hypotheses (+3 by agreement with the orchestrator → max 8) | escalation to the user (escalated_to_user) |
| BLOCK iterations of artifact review | reviewer / orchestrator | 3 iterations | escalation to the orchestrator/user |
| Review of the debugger's local fix (scope=debug) | reviewer / orchestrator | 1 iteration | escalation to the orchestrator/user |
| bug → fix → bug again cycle on the same artifact | orchestrator | 2 cycles | escalation to the orchestrator/user |
depends_on: []
Related Skills
steelmorgan/onec-server-maintenance-hooks
development
VerifiedTrustedCommunity
1C server maintenance webhooks: container restart and external component cache cleanup
83SKILL.mdUpdated Jul 4, 2026
steelmorgan/dap-bsl-code-debug-procedure
development
VerifiedTrustedCommunity
Interactive DAP debugging of a single BSL procedure
83SKILL.mdUpdated Jul 4, 2026
steelmorgan/rlm-bsl-search
tools
VerifiedTrustedCommunity
Rules for using RLM tools for project search and navigation in 1C/BSL
83SKILL.mdUpdated Jul 4, 2026
steelmorgan/winow
development
VerifiedTrustedCommunity
Creates web applications and routes on Winow (a web server on OneScript and Autumn). Use when working with a web server on OneScript, routing, or Winow controllers.
83SKILL.mdUpdated Jul 4, 2026