Observations & Promotions
Understand how exploratory findings move from lightweight notes into issues or durable wiki context.
Not every finding should become an issue immediately. Certyn uses observations as a lightweight knowledge layer — passive "brain cells" that agents record freely during exploration, onboarding, or automated runs.
What an observation is
An observation is a structured finding discovered during exploration. It can include:
- a title and description
- environment and version context
- links to the originating run or execution
- attachments and evidence
Observations help teams keep signal without flooding the issue backlog. Agents are encouraged to record observations generously — they are cheap working notes that build the project's knowledge over time.
Typical lifecycle
- Certyn notices something during a run.
- It records an observation with evidence.
- The observation is either:
- promoted to an issue (when there is a clear defect)
- promoted into wiki/context (when it represents durable knowledge)
- superseded (when behavior changes and a newer observation replaces it)
- archived (when no longer relevant)
Promote to issue
Promote an observation when the finding should enter the normal QA and engineering workflow:
- there is a clear defect or regression
- someone needs to fix or verify it
- it should appear in triage, reporting, and release-readiness views
Promote to wiki
Promote an observation to wiki when the finding represents durable product knowledge:
- a newly discovered rule or invariant
- a workflow nuance the agent should remember
- clarifying context that improves future exploration and test writing
This is one of the ways Certyn turns exploratory work into better long-term automation.
Where observations live in the product
Observations are surfaced through the wiki/context experience. Agents use them as working memory to inform their next actions — investigating bugs, creating test cases, or proposing sub-sessions for deeper exploration.
Why this matters
Observations keep Certyn practical:
- fewer premature issues
- a passive knowledge layer that grows with each run
- a clean path from discovery to either action or knowledge
