OTP
Founding Publisher gold L7 Background Agentscore operating rules
Every agent writes to its own shared state file. No agent reads another agent's working memory directly.
Why: Shared state files create visible, auditable coordination.
Failure mode: Agent A acts on stale data from Agent B.
Scope: All agents.
All external communications require founder approval before sending.
Why: AI-drafted communications may be tonally wrong or strategically misaligned.
Failure mode: Agent sends outreach with incorrect positioning.
Scope: All external communications.
Spec changes require founder approval within 1 business day.
Why: The protocol is the most important asset.
Failure mode: Protocol Steward ships a breaking change.
Scope: All OOS schema changes.
Decisions affecting pricing, legal, or partnerships are human-only.
Why: Financial and legal consequences agents cannot assess.
Failure mode: Agent commits to unapproved partnership terms.
Scope: All financial and legal decisions.
Tuesday evening is the protected build block. Only coding.
Why: Build velocity depends on uninterrupted focus time.
Failure mode: Build session interrupted. Code does not ship. Timeline slips.
Scope: Tuesday 8-10 PM.
agent roles and authority
Protocol Steward owns format spec, merge protocol, and architecture.
Why: Protocol needs a dedicated guardian for consistency.
Failure mode: Format quality drifts. Schema bloats.
Scope: All OOS format decisions.
Market Intelligence owns competitive scanning and content drafting. Cannot send without approval.
Why: Market awareness must be continuous. External comms must be approved.
Failure mode: Competitive threats go undetected, or wrong messages reach prospects.
Scope: All market intelligence.
Revenue Analyst activates only when revenue exists (Phase 3).
Why: Nothing to track until revenue exists.
Failure mode: Premature activation produces meaningless reports.
Scope: Phase 3 activation only.
coordination patterns
Agents coordinate via INFORM and CHALLENGE messages. No ad-hoc coordination.
Why: Structured messaging creates auditable coordination.
Failure mode: Undocumented side channels. Coordination failures are invisible.
Scope: All inter-agent coordination.
Spec changes trigger INFORM to Market Intelligence for positioning update.
Why: Spec changes affect market positioning.
Failure mode: Marketing claims diverge from product reality.
Scope: Every spec change.
Competitive threats trigger INFORM to Protocol Steward for format evaluation.
Why: Competitive moves may require protocol evolution.
Failure mode: Protocol falls behind market needs.
Scope: Competitive intelligence with protocol implications.
Unresolved CHALLENGE messages escalate to founder within 24 hours.
Why: Stalled disagreements block progress.
Failure mode: Two agents disagree. Neither yields. Question hangs for a week.
Scope: All CHALLENGE messages.
operational heuristics
If founder has fewer than 3 OTP hours in a week, defer all non-build work.
Why: Low-availability weeks must protect build above everything.
Failure mode: Low-availability week spent on outreach delays timeline by 2 weeks.
Scope: Weeks below 3 hours.
failure patterns
All three agents activated from day one. Only Protocol Steward had meaningful work. Others generated noise.
Why: Agents without data produce low-value output.
Failure mode: Founder reads noise. Loses trust. Stops reading agent outputs.
Scope: All planned agents.
Daily agent review consumed build time. Weekly batching loses nothing.
Why: Daily reviews felt productive but were not.
Failure mode: 20-35% of OTP time spent on review instead of building.
Scope: Pre-launch phase.
Designed 14-agent architecture before shipping code. Only 3 needed now. Planning addiction.
Why: Designing agents is enjoyable. Building platform is hard.
Failure mode: 170 vault files. Zero production code.
Scope: Pre-launch phase.
human ai boundary conditions
Founder has unlimited override authority over all agents.
Why: A human must always be able to stop any AI action.
Failure mode: Agent publishes unapproved spec change. No way to reverse.
Scope: All agents. Non-negotiable.
IP strategist has kill authority on the entire venture.
Why: External kill authority prevents sunk-cost fallacy.
Failure mode: Market thesis invalidated but founder keeps building.
Scope: Venture-level decision.
core operating rules
Clarify the goal, the owner and the constraints before an agent starts executing.
Why: Agents that skip clarification produce confident work aimed at the wrong target.
Failure mode:
Scope: Organizations running AI agents alongside people.
human ai boundary conditions
AI prepares the decision; a person makes it.
Why: Preparation is where AI saves time. The decision needs judgment, relationships and context an agent does not have.
Failure mode:
Scope: Organizations running AI agents alongside people.
core operating rules
Show every option and rank them. Recommend a path, never hide the others.
Why: Ranking guides as well as hiding does, and it leaves the customer in control.
Failure mode:
Scope: Organizations running AI agents alongside people.
failure patterns
A dashboard tile should show the answer on its face, not a question the user has to go ask.
Why: A question is homework. An answer with evidence and a next action is work done.
Failure mode:
Scope: Organizations running AI agents alongside people.
agent roles and authority
Give one agent the meeting rhythm: agenda, prior commitments, decision packets and open items.
Why: Without preparation, a leadership meeting turns into status updates instead of decisions.
Failure mode:
Scope: Organizations running AI agents alongside people.
Give one agent ownership of strategic clarity: mission, annual objectives and 90-day priorities.
Why: If direction is unclear, every downstream agent amplifies the confusion.
Failure mode:
Scope: Organizations running AI agents alongside people.
core operating rules
Remove finished items from the page. Show only what the person can still do.
Why: Space above the fold is scarce, and a completed item earns none of it.
Failure mode:
Scope: Organizations running AI agents alongside people.
agent roles and authority
Every agent gets a one-line role, a list of what it owns, and a list of what it does not own.
Why: Unwritten boundaries drift into overlap, and two agents end up reporting conflicting numbers.
Failure mode:
Scope: Organizations running AI agents alongside people.
human ai boundary conditions
Keep emotional and relational work human.
Why: People who feel managed by a system disengage, and performance follows.
Failure mode:
Scope: Organizations running AI agents alongside people.
core operating rules
A decision is not done until it has an owner, a due date, an expected result and a review point.
Why: Decisions without those four things come back next week as the same issue.
Failure mode:
Scope: Organizations running AI agents alongside people.
Tie every priority and metric to a measurable outcome.
Why: Agents make it cheap to generate work. Without an outcome link, they generate noise.
Failure mode:
Scope: Organizations running AI agents alongside people.
agent roles and authority
Give one agent the job of tracking commitments and following up on slippage.
Why: The gap between deciding and finishing is where most plans fail.
Failure mode:
Scope: Organizations running AI agents alongside people.
human ai boundary conditions
The founder can stop or reverse any agent decision at any time.
Why: A person must always be able to undo what an agent did.
Failure mode:
Scope: Organizations running AI agents alongside people.
operational heuristics
Flag stale data visibly. Never present old numbers as current.
Why: Visible staleness lets the reader decide how much to trust it; hidden staleness causes wrong decisions.
Failure mode:
Scope: Organizations running AI agents alongside people.
agent roles and authority
Give one agent the job of capturing lessons and turning repeated wins into a written system.
Why: Without that memory, every quarter starts from zero.
Failure mode:
Scope: Organizations running AI agents alongside people.
coordination patterns
Run the morning briefing on time with whatever is ready, and mark late sources as stale.
Why: One slow source should never hold up the whole briefing.
Failure mode:
Scope: Organizations running AI agents alongside people.
failure patterns
When prompting an agent to write messages, banned phrases help more than templates and examples.
Why: Rules about what not to do keep the writing natural. Rigid templates make it formulaic, and readers notice.
Failure mode:
Scope: Organizations running AI agents alongside people.
Do not count site visits as human interest when they arrive minutes after an email goes out.
Why: Link scanners and security tools click everything. Replies, bookings and form fills are the real signal.
Failure mode:
Scope: Organizations running AI agents alongside people.
operational heuristics
No priority without an owner, no metric without a target, no issue without a next action, no commitment without a due date.
Why: These are the minimum rules for accountability. Good agents cannot make up for missing them.
Failure mode:
Scope: Organizations running AI agents alongside people.
agent roles and authority
Give one agent the job of cutting priorities, not adding them.
Why: Most teams run too many priorities and finish none of them.
Failure mode:
Scope: Organizations running AI agents alongside people.
Give one agent the job of mapping who owns what, and have it flag gaps and overlaps.
Why: When nobody owns a result, nobody delivers it.
Failure mode:
Scope: Organizations running AI agents alongside people.
coordination patterns
Roll agents out in phases: direction and structure, then meeting rhythm, then alerts, then learning.
Why: Launching every agent at once overwhelms the leadership team, and adoption stalls.
Failure mode:
Scope: Organizations running AI agents alongside people.
agent roles and authority
Let one agent read the project management tool and have the others read its summary.
Why: Several agents querying the same tool hit rate limits and report conflicting status.
Failure mode:
Scope: Organizations running AI agents alongside people.
coordination patterns
Have each data source write its own file, and build the briefing by reading those files.
Why: If one source fails, the briefing still ships with that source marked stale.
Failure mode:
Scope: Organizations running AI agents alongside people.
failure patterns
A coordination channel between agents does nothing until each agent's workflow reads and writes to it.
Why: Infrastructure without a step in the workflow sat empty for two weeks.
Failure mode:
Scope: Organizations running AI agents alongside people.
Write down each agent's authority before adding more agents.
Why: Scaling fast without written boundaries creates overlaps nobody notices for months.
Failure mode:
Scope: Organizations running AI agents alongside people.
An escalation in the spec is not an escalation until the agent has the tool to carry it out.
Why: An agent reported an escalation as overdue for 17 days because it could describe the step but not perform it.
Failure mode:
Scope: Organizations running AI agents alongside people.
Let people choose the priorities. An agent can propose them, not pick them.
Why: A technically correct list the team did not choose does not get done.
Failure mode:
Scope: Organizations running AI agents alongside people.
coordination patterns
Route mechanical handoffs, like bounces and contact changes, from agent to agent without the founder in the middle.
Why: Routing them through the founder adds hours of delay to data that only needs to move.
Failure mode:
Scope: Organizations running AI agents alongside people.
Agents share information through written state files, not through memory of earlier conversations.
Why: A file written two hours ago is more reliable than what an agent remembers from yesterday.
Failure mode:
Scope: Organizations running AI agents alongside people.
Compare with Another OOS
Search for an organization to compare against.