Use when running end-to-end STM32 debug campaigns that coordinate flash, RTT, register capture, HardFault snapshots, and error archival across multiple tools.
BFD-Kit/.learnings/ERRORS.md and/or BFD-Kit/.learnings/LEARNINGS.md; unresolved capability gaps go to BFD-Kit/.learnings/FEATURE_REQUESTS.md.SKILL.md, script, wrapper, or resource so the next run benefits by default.BFD-Kit/.learnings/CHANGELOG.md and mention the improvement in the task close-out.Use this skill to coordinate multi-step debug campaigns with profile-driven device settings.
bfd-project-initbfd-flash-programmerbfd-rtt-loggerbfd-debug-interfacebfd-debug-executorbfd-register-capturebfd-fault-logger# 0) Bootstrap profile (required)
python3 ./.claude/skills/bfd-project-init/scripts/bootstrap.py --project-root . --mode check
# Full fault campaign
./.claude/skills/bfd-debug-orchestrator/scripts/run_fault_campaign.sh
# Inject one scenario
./.claude/skills/bfd-debug-orchestrator/scripts/inject_fault_scenario.sh --scenario 3
# Capture one HardFault snapshot
./.claude/skills/bfd-debug-orchestrator/scripts/capture_hardfault_snapshot.sh
# Manual quick RTT capture without reset (attach-only, for soft fault 1/2)
./build_tools/jlink/rtt.sh logs/rtt/manual_quick.log 4 --mode quick
# Manual dual RTT capture after recovery reset/go
./build_tools/jlink/rtt.sh logs/rtt/manual_dual.log 6 --mode dual --reset-policy gdb-reset-go
1: recoverable IMU communication fault2: recoverable Flash parameter fault3: illegal address write (HardFault)4: UDF trap (UsageFault/HardFault)quick: attach-only RTT capture for scenarios 1/2; do not issue reset before collecting the injected soft-fault log.dual: reset-aware recovery capture for scenarios 3/4; use GDBServer + RTTClient and let GDB own the reset/go sequence.quick, but those calls are no longer allowed to depend on build_tools/jlink/rtt.jlink.md and json records.logs/ or .claude/debug/.quick mode is attach-only and must not reset away g_debug_fault_scenario before soft-fault evidence is captured.dual mode is reserved for reset-aware recovery verification and should drive reset/go through the same GDB server backend..claude/skills/bfd-debug-orchestrator/references/hardfault_record_template.md.claude/skills/bfd-debug-orchestrator/references/error_evolution_schema.md.claude/skills/bfd-debug-orchestrator/references/token_saving_output_rules.md.claude/skills/bfd-debug-orchestrator/resources/README.md.claude/skills/bfd-debug-orchestrator/resources/f4/debug_fault_template.h.claude/skills/bfd-debug-orchestrator/resources/f4/debug_fault_template.c