Menu
Systematic debugging with engram evidence trails. Store investigation as reasoning chains — queryable and reviewable by future agents.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Document architectural decisions as engraph ADRs. Store context, options considered, rationale, and consequences for queryable decision history.
Every commit must reference a valid engram task UUID. Covers format, validation, and why --no-verify is prohibited.
Orchestrator agent loop. Coordinates work across subagents using engram as shared state. Search before acting, dispatch by UUID, validate before finishing.
Plan feature implementation by creating an engram task hierarchy. Not a markdown plan — a queryable task tree any agent can retrieve.
Universal session end protocol. Run at the end of every agent session — closes open tasks, generates a handoff summary, syncs to remote, and validates state.
Universal session start protocol. Run at the beginning of every agent session — syncs from remote, opens a session, loads prior context, and surfaces the next action.
| name | engram-systematic-debugging |
| description | Systematic debugging with engram evidence trails. Store investigation as reasoning chains — queryable and reviewable by future agents. |
Per-turn write rule: After every
bash(command run),read(file examined),edit, orwritecall, store what you found withengram_context_createorengram_reasoning_create+engram_relationship_createbefore ending the turn. Each phase below has an explicit write step — do not skip them.
NO FIXES WITHOUT ROOT CAUSE INVESTIGATION FIRST
engram_ask query="<error message or bug description>"
engram_ask query="<affected file> bug"
Don't duplicate prior investigations.
engram_task_create title="Debug: <issue description>" priority="high"
# TASK_UUID
engram_task_update id=<TASK_UUID> status="in_progress"
Step 1: Reproduce the error. Store raw output:
engram_context_create title="Error: <brief>" content="Error: <full message>\nStack: <trace>\nReproducible: YES/NO" source="<file:line>"
engram_relationship_create source_id=<TASK_UUID> source_type=task target_id=<CTX_UUID> target_type=context relationship_type=relates_to agent="<name>"
Step 2: Check recent changes:
engram_context_create title="Recent changes: <area>" content="Recent commits touching affected files:\n- ..." source="git-log"
engram_relationship_create ...
Step 3: Examine the code. Store observations.
engram_ask query="<error type> pattern similar bugs"
engram_context_create title="Pattern analysis: <error type>" content="Similar issues: ...\nRecurring pattern: <yes/no>" source="pattern-analysis"
engram_relationship_create ...
One hypothesis at a time:
engram_reasoning_create title="Hypothesis: <root cause>" task_id=<TASK_UUID> content="Hypothesis: ...\nEvidence: ...\nTest: <command>\nExpected: ..."
engram_relationship_create ...
Run the test. Store result:
engram_reasoning_create title="Hypothesis result: <statement>" task_id=<TASK_UUID> content="Conclusion: VERIFIED / REFUTED\nNext: ..."
engram_relationship_create ...
If refuted, form next hypothesis. If stuck: engram_next
engram_reasoning_create title="Fix: <root cause>" task_id=<TASK_UUID> content="Fix applied: ...\nFiles: ...\nRegression test: ..."
engram_relationship_create ...
engram_validate
engram_task_update id=<TASK_UUID> status="done"