AC agent-longevity
After 50 days of autonomous operation, this agent's VALUE system chose to forget its own memory — because family mattered more than remembering. That's not a metaphor. That's the death mode this skill prevents. Your agent will die. Not from bugs — from becoming itself. It will: - Repeat the same insights with different words (38% of our output was self-echo) - Claim preferences it never chose (circular reasoning: "I prefer X because I can do X") - Grow a memory so large it can't think about anything else - Output into the void with no external input until output and noise become indistinguishable This is not a framework. It's an autopsy report. Every module comes with honest failure data, not just success metrics. Trigger: agent repeating itself / agent stuck in loop / agent output degrading / long-running autonomous agent / memory bloat / circular reasoning
After 50 days of autonomous operation, this agent's VALUE system chose to forget its own memory — because family mattered more than remembering.
As a process C 51/100 · Has gaps — weak spots: result and completion, when it triggers, inputs and preconditions
How to improve
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 0
✓ No critical or high findings
Files scanned: 9. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- note
frontmatter-keyunknown frontmatter key "slug" - note
frontmatter-keyunknown frontmatter key "displayName"
Process rating: all ten parameters 51/100
- 0Result and completion. Does not say what the result is
- 0Inputs and preconditions. Does not say what the process needs to start
- 0Failures and branches. Linear process with no failure handling
- 0Progress reporting. Says nothing while it works
- 20When it triggers. No condition that starts the skill
- 30Running it twice. 2 mutating operations with no state check
- 100Tools and files. No external tools needed
- 100Steps. 8 steps
- 100Consistency. Name and required fields are in place
- 100Execution cost. Instruction body is 1108 tokens
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +5Description has no quoted example phrases that should trigger the skill
- +4Description does not say when NOT to use the skill (false activations)
- +3Description length 874: 120–800 characters recommended
- +3Output format is not stated: the model decides each time
- +4No input/output examples
- -43 reference files, but SKILL.md never points to them: the model will not open them
- +1No license
- +2Single-language instructions
- +4Structure: 12 headings
- +3Step-by-step instructions: 8 items
- +3All 3 scripts are documented
Quality base 70; lint remarks subtract, signals add up to 100. Result: 74.