# Governance Is the Platform Problem Worth Solving > [!question] Lingering Questions > - Who should hold an agent's identity when agents chain to other agents — the invoking human, the invoking agent, or a freshly scoped-down identity minted per call? > - What does an "agent audit trail" schema need to contain to be useful during a live incident, not just after the fact? > - At what organizational size or regulatory exposure does a federated (team-based) governance approach stop working and a centralized model become necessary? > [!tip] Article Ideas > - The EU AI Act vs. Canadian government AI guidance contrast (bureaucratic-detailed vs. simple-checklist) as a short comparison piece for the governance wiki. #articleseed > [!link] Cross Reference > - [[3. Resources/Permanent Notes/Identity Is the Control Plane in Distributed Systems]] — already captures the agent-identity-delegation claim from this exact source (already listed in that note's Source section) > - [[3. Resources/Permanent Notes/A Watchable Pilot Earns Buy-In Faster Than ROI Proof]] — adjacent buy-in territory, different mechanism (watchability vs. KPI-translation) > - [[3. Resources/MOCs/AI Agent Safety MOC]] > - [[3. Resources/MOCs/AI & the Future of Knowledge Work MOC]] > [!quote] Vocabulary > - **Agent identity delegation** — the unresolved question of whose identity/permissions an agent inherits when it invokes another agent in a chain. > - **Theory of change** — mapping inputs (cost, time, data) through activities to outcomes, used here as an ROI-justification framework when the causal link is indirect. > - **Golden path** — the platform team's supported, opinionated default workflow; referenced regarding why agents are kept out of internal developer platforms during low-trust phases. ## Bib capsule LeadDev (webinar produced with Harness), "Governance is the platform problem worth solving," YouTube, August 13, 2026. Panelists: Matali (Yelp), Mark Boyd (Platformable), Pratik Mutal (Harness). https://www.youtube.com/watch?v=Xo2MCXQtiyI ## Author's thesis (Q1+Q2) AI agent governance is best treated as an extension of existing organizational governance disciplines (security, data, API) rather than a wholly new one. The practical work is defining agent identity and access scope, building agent-level observability/tracing, securing cross-functional buy-in by translating governance benefits into decision-makers' existing KPIs, and publishing an imperfect policy now rather than waiting for a perfect one. ## Key arguments (≤3) 1. **Governance disciplines converge on "context."** Data, API, and AI governance all cash out as describing who is using something, how, and under what guardrails, at the right level of granularity. 2. **Agent-to-agent identity delegation is the sharpest unresolved risk.** A technically sound governance policy can still be defeated by an over-permissioned agent chain — "the agent can go rogue... break out of that governance model." 3. **Governance buy-in is an organizational-design problem before a technical one.** It requires knowing the org's structure, the actual decision-makers, what those decision-makers' KPIs already are, and translating governance's value into that language. ## Claim atoms ### Trust is the core unit of governed ecosystems > [!quote] ID:rw_gov01 > the core unit that makes an open digital ecosystem work is trust Where/why: Mark Boyd, ~06:14, answering "what is AI agent governance vs API governance" — anchors the whole panel's framing that governance exists to produce trust, not satisfy a compliance checklist. ### Unscoped agent delegation lets agents go rogue > [!quote] ID:rw_gov02 > the agent can go rogue and then it can actually... break out of that governance model Where/why: Pratik Mutal, ~15:20, on agent-to-agent identity delegation — names the sharpest concrete risk in the discussion. ### Yelp merged security into developer experience > [!quote] ID:rw_gov03 > it was actually just last year that we reorged to have security in the same bigger organization as developer experience Where/why: Matali, ~27:40, describing the org-structure change behind Yelp's cross-functional buy-in — a concrete, replicable structural pattern. ## Permanent candidates - The agent-identity-delegation risk (H2) is already captured by [[3. Resources/Permanent Notes/Identity Is the Control Plane in Distributed Systems]], which already lists this exact source. No duplicate created. - **New**: the governance-buy-in-follows-KPIs claim (H1/P1 in the source deep-read) is not yet captured — promoted below as [[3. Resources/Permanent Notes/Governance Funding Follows the Decision-Makers Existing KPIs]]. ## Processing checklist - [x] Checked for existing Literature Note — none found - [x] Checked for existing Permanent Notes covering this source's claims — found [[3. Resources/Permanent Notes/Identity Is the Control Plane in Distributed Systems]] already covering identity delegation and already sourcing this doc; did not duplicate - [x] Connected into [[3. Resources/MOCs/AI Agent Safety MOC]] and [[3. Resources/MOCs/AI & the Future of Knowledge Work MOC]] ## Links - [[3. Resources/MOCs/AI Agent Safety MOC]] — identity/access-scoping and observability-for-agents themes - [[3. Resources/MOCs/AI & the Future of Knowledge Work MOC]] — governance-as-platform-trust and ROI-measurement themes - [[3. Resources/Permanent Notes/Governance Funding Follows the Decision-Makers Existing KPIs]] — extends: the permanent note promoted from the panel's buy-in argument - [[3. Resources/Permanent Notes/Identity Is the Control Plane in Distributed Systems]] — extends: carries the agent-identity-delegation risk the panel calls the sharpest open problem - [[3. Resources/Permanent Notes/An Identity Model That Predates Delegation Cannot Attribute Delegated Action]] — evidence-for: why an agent chain defeats a governance policy when audit only sees the borrowed credential - [[3. Resources/Permanent Notes/A Watchable Pilot Earns Buy-In Faster Than ROI Proof]] — contrasts: the panel translates governance into KPIs, this note argues a visible pilot builds the proof instead