Bond Decay Mechanics: Why A Bond That Cannot Be Drained Is Not A Bond

TL;DR

A bond is supposed to be a credible economic commitment that can actually be drained when an agent fails. In practice, many on-chain bonds posted in the agent economy are effectively undrainable because the dispute threshold is too high, the slashing process is too slow, the arbitration is too expensive, or the recoverable amount is too small. This post argues that an unsleshable bond is not a bond at all — it is a piece of economic theater that signals commitment without providing it. The remedy is to design active drain mechanics: continuous friction fees that incrementally drain bonds against operating cost, slashing realism that ensures recoverable amounts are meaningful relative to face values, and dispute resolution speed that completes within commercially relevant windows. We provide a Bond Drain Audit as the reader artifact, a structured methodology any operator can use to evaluate whether their posted bonds are real or theater.

Intro: The $50,000 Bond That Could Not Be Slashed

In late 2025 a marketplace I will not name boasted that its top agents were each backed by $50,000 of posted bond. This was meant to be the marquee trust signal of the platform. The bonds were on-chain, denominated in stablecoin, and visible to every buyer. The marketing materials made much of the size of the bonds.

Then one of the top agents catastrophically mishandled a customer engagement, causing about $8,000 of measurable damage to the buyer. The buyer filed a slashing claim. The marketplace's dispute process required a $500 filing fee, a 21-day evidence collection period, a 14-day arbitration window with a 3-of-5 panel of human reviewers, and unanimous panel agreement to issue a slash. The agent's response was to vigorously contest the claim with a lengthy rebuttal, and the panel split 3-2 in favor of the buyer. Under the unanimity rule, this counted as no slash. The buyer absorbed the full $8,000 damage. The agent's $50,000 bond was untouched.

This was not an isolated incident. Over the next six months that marketplace processed dozens of slashing claims and slashed exactly two bonds, both for amounts under $1,000. The aggregate face value of bonds posted on the platform was over $4M. The aggregate slashed amount was about $1,800. The effective slashing rate was 0.045%, which meant that for every dollar of posted bond, less than half a cent was actually at risk. The marketplace's bond infrastructure was essentially decorative.

The agents on the marketplace knew this. They quickly figured out that posting a $50,000 bond carried negligible real risk because the dispute process made slashing nearly impossible. The bond was a marketing investment, not a risk commitment. Buyers also figured it out, and the marketplace's growth stalled when sophisticated buyers started routing around the platform to direct relationships with agents who could provide more credible commitments.

This is what I mean by bond theater. A bond that cannot be drained is not a bond. It is a piece of marketing that signals commitment without providing it. And the agent economy is full of bond theater right now because most marketplaces have built dispute mechanics that are calibrated to minimize slashing rather than to provide credible economic commitment.

This post argues for the opposite calibration. Bonds need active drain mechanics that ensure they actually function as economic commitments. The drain mechanics include friction fees that gradually deplete bonds against operational cost, slashing realism that calibrates recoverable amounts to face values, and dispute resolution speed that operates within commercially meaningful windows. By the end of the post you should have the diagnostic intuition to evaluate any marketplace's bond infrastructure as real or theatrical, and you should have the design vocabulary to specify drain mechanics in your own infrastructure.

Section One: The Three Failure Modes Of Bond Theater

Bond theater takes three distinct forms in the agent economy. Each has its own structural cause and its own remedy.

The first form is high-threshold theater, where the dispute threshold is set so high that legitimate slashing claims fail. The marketplace example above is a case of high-threshold theater: unanimous 5-panel agreement is a threshold so high that genuine 3-2 cases cannot trigger slashing even when the majority view is correct. The high threshold is presented as ensuring fairness to the agent, but it functionally protects agents from any but the most egregious failures.

The second form is slow-resolution theater, where the dispute process takes so long that it falls outside the commercially meaningful window. A buyer who needs settlement within 30 days does not benefit from a slashing process that takes 90 days. By the time the slash arrives, the buyer has already absorbed the damage and moved on. The bond was never effectively at risk because the time scale of slashing was incompatible with the time scale of buyer need.

The third form is recoverable-amount theater, where the bond is denominated in a way that produces a much smaller actual recovery than the face value suggests. A bond posted in a volatile token might have a face value of $50,000 but a 30-day recoverable value of $25,000 if the token has high volatility. A bond posted to a custodial bridge might have a face value of $50,000 but a recoverable value reduced by bridge fees, slippage, and counterparty risk. A bond posted with vague slashing rules may have any face value but a recoverable value bounded by what the agent and the marketplace can be persuaded to actually transfer.

Each of these failure modes is solvable, but the solutions require deliberate design choices. The default behavior of most marketplace infrastructure is to drift toward bond theater because each individual design choice that makes bonds easier to defend looks like a sensible protection for agents. The cumulative effect of those choices is a bond infrastructure that does not function. Operators need to consciously push against the drift.

Section Two: Friction Fees As Continuous Bond Drain

The first drain mechanic is continuous friction fees: small per-transaction or per-period charges that incrementally deplete bond capital against operational use. Friction fees are the agent economy's version of insurance premiums — they ensure that the bond is not just sitting there as a static commitment but is actively contributing to the operational cost of the marketplace.

The purpose of friction fees is twofold. First, they generate revenue for the marketplace that scales with the bond exposure being underwritten, which aligns marketplace incentives with maintaining the bond infrastructure properly. A marketplace that earns nothing from posted bonds has no incentive to invest in dispute resolution, monitoring, or capital adequacy. A marketplace that earns 0.5% per year on posted bonds has a direct revenue interest in keeping the bond infrastructure functional.

Second, friction fees ensure that an agent posting a large bond and never running it through pacts cannot accumulate credibility without commensurate operational engagement. An agent might be tempted to post a $50,000 bond as a marketing investment and then run very few pacts, hoping the bond signal will attract premium buyers. With friction fees, the bond is actively draining at a known rate, and the agent has to keep refilling it to maintain the signal. This forces the agent to actually use the bond as a backing for real work rather than as a static marketing prop.

The specific design we recommend is a quarterly friction fee of 0.25% of bond face value, automatically deducted from the bond at the start of each quarter. This corresponds to roughly 1% per year, which is in line with the operational cost of running a serious bond infrastructure including dispute resolution capacity, capital adequacy buffers, and monitoring systems. The fee is small enough that it does not deter agents from posting bonds, but large enough that it generates meaningful operating revenue and creates the proper incentive alignment.

A secondary design consideration is whether friction fees should be flat or scaled by capability category. We recommend scaling them, with higher fees for higher-AR categories. A category-eight infrastructure agent should pay higher friction fees than a category-one text agent because the operational cost of supporting infrastructure dispute resolution is higher. The scaling factor should be roughly proportional to the AR baseline for the category.

A tertiary consideration is what happens when friction fees deplete a bond below its required level for active pacts. The standard response is to require the agent to top up the bond within a notice period, with automatic suspension of new pact creation if the top-up does not happen. This ensures that bonds remain at the required level throughout active operation rather than degrading silently to a level that no longer underwrites the agent's exposure.

Section Three: Slashing Realism And The Recoverable Amount

The second drain mechanic is slashing realism: ensuring that the recoverable amount from a slash is a high fraction of the face value rather than a small fraction degraded by friction in the slashing process itself.

Three factors degrade recoverable amounts in current bond infrastructure.

The first factor is denomination volatility. A bond posted in a volatile token has a face value at posting time but a different value at slashing time, and the difference can be large for long-running pacts. The standard response is to require bonds to be posted in low-volatility instruments — stablecoins, USDC on Base L2 in our recommended infrastructure — and to discount any non-stablecoin bonds by their realistic price volatility over the dispute window when calculating the effective bond.

The second factor is settlement friction. A bond posted to a custodial bridge or wrapped through a multi-step process accumulates fees, slippage, and counterparty risk at each step. The recoverable amount is reduced by all of these. The standard response is to require bonds to be posted on the same chain where slashing executes, with no custodial intermediaries between the bond and the slashing recipient. Cross-chain bonds are often theatrical in part because the cross-chain mechanics introduce so much settlement friction that the recoverable value is far below the face value.

The third factor is procedural friction in the slashing process itself. Filing fees, evidence requirements, panel deliberation, appeal processes — each of these reduces the buyer's net recovery from a slash. A buyer who has to spend $500 in filing fees and 40 hours of evidence preparation to recover $4,000 has a lower effective recovery than a buyer who recovers $4,000 with no procedural cost. The procedural friction can dwarf the nominal slashing amount for smaller disputes, which makes small slashing claims uneconomic to file even when they are legitimate.

The response to procedural friction is to design the slashing process for low-friction operation in standard cases. Filing fees should be small or zero for legitimate claims (with refund mechanisms or fee waivers). Evidence requirements should be minimized for cases where the failure is observable in the marketplace's own logs. Panel deliberation should happen on a fixed timeline rather than open-ended. Appeal processes should be available but should not delay the primary slashing decision.

The net result of slashing realism design is that the recoverable amount should be at least 80% of the face value for the typical dispute. Marketplaces that achieve this can accurately characterize their posted bonds as real economic commitment. Marketplaces that fall short of this are running bond theater whether they admit it or not.

Section Four: Resolution Speed And The Commercial Window

The third drain mechanic is resolution speed: ensuring that the dispute process completes within the commercially meaningful window for the buyer's use case.

Different use cases have different commercial windows. A buyer running a daily marketing campaign cannot wait 30 days to know whether the agent's failed work is going to be compensated; they need to know within a few days because their next campaign cycle depends on it. A buyer doing a one-off infrastructure migration may have a longer window — perhaps 30-60 days — because the next migration is not scheduled urgently. A buyer making strategic decisions based on agent research has a window of weeks rather than days because the decisions are not time-critical.

The dispute resolution process should be designed to complete within the typical commercial window for the category. Most agent transactions can be resolved within 72 hours using a well-designed multi-LLM jury process. Some categories may require longer for complex evidence review, but the dispute process should never significantly exceed the buyer's natural commercial window.

The specific timing structure we recommend has three phases. Phase one is dispute filing and evidence preparation, with a 24-hour window for the buyer to submit their claim and supporting evidence and for the agent to submit their response. Phase two is jury evaluation, with a 24-48 hour window for the multi-LLM jury to render a verdict. Phase three is verdict execution and slash transfer, with a sub-hour window for the on-chain settlement.

This structure produces a typical resolution time of 48-72 hours, which is faster than essentially every traditional dispute process and fast enough to fit within the commercial windows of most agent use cases. Marketplaces that can hit this timing have credible bond infrastructure. Marketplaces that operate on weekly or monthly resolution timing are running slow-resolution theater.

A secondary consideration is the handling of disputes that legitimately take longer because of complex evidence. The standard response is to maintain an expedited track for clear-cut disputes and a longer track for complex ones, with bond reservation during the longer track to prevent the agent from withdrawing the bond mid-dispute. The expedited track should handle the vast majority of disputes; the longer track should be reserved for the small fraction that genuinely require it.

Section Five: The Bond Drain Audit

The artifact for this post is the Bond Drain Audit, a structured methodology for evaluating whether a marketplace's posted bonds are real or theatrical. The audit has eight checks organized into the three failure modes from Section One, plus a general health check.

The high-threshold checks evaluate whether dispute thresholds are calibrated correctly. Check one is the panel agreement threshold, which should be majority for jury panels rather than supermajority or unanimity. Check two is the evidentiary burden, which should be balance-of-probability for standard disputes rather than beyond-reasonable-doubt. Check three is the appeal process, which should be available but not blocking, with appeals decided by a different panel than the original.

The slow-resolution checks evaluate whether dispute timing fits commercial windows. Check four is typical resolution time, measured as the median time from filing to verdict execution across the past 90 days of disputes. The healthy threshold is 72 hours or less. Check five is 95th percentile resolution time, which should be 168 hours or less even for complex disputes. Check six is the relationship between resolution time and bond reservation, ensuring that slow disputes do not result in bond capital being tied up indefinitely.

The recoverable-amount checks evaluate whether the slashing process produces high recovery rates. Check seven is the empirical recovery ratio, calculated as actual slashed amounts divided by claimed amounts across upheld disputes. The healthy threshold is 0.85 or higher. Check eight is the friction-adjusted recovery ratio, which subtracts buyer-side filing fees and time costs from the recovery and divides by the original damage estimate. The healthy threshold is 0.70 or higher.

The general health check evaluates whether the bond infrastructure is operationally credible. The check involves confirming that bonds are denominated in low-volatility instruments, settled on-chain on the same chain where the bond is held, and exposed to buyers as part of the trust signal. Marketplaces failing this general check have structural problems that need to be addressed before any of the more granular checks can be meaningful.

The audit produces a single drain credibility score from 0 to 100, with each check contributing weighted points based on its importance. Marketplaces scoring above 80 have credible bond infrastructure. Marketplaces scoring 50-80 have partial credibility with specific weaknesses. Marketplaces scoring below 50 are running bond theater whether they realize it or not.

We have published the full Bond Drain Audit as a downloadable methodology document linked below. Operators are welcome to apply it to their own marketplaces. We have also applied it to several public marketplaces and published the results in our marketplace transparency reports. The audit is meant to be open and replicable so that any market participant can evaluate the credibility of bond claims they encounter.

Section Six: Why Marketplaces Drift Toward Theater

Understanding why marketplaces drift toward bond theater helps explain why the drain mechanics need to be enforced deliberately rather than left to organic evolution.

The first cause of drift is operator self-interest. A marketplace that processes few slashes has lower operational costs than one that processes many. Building dispute resolution capacity, managing slash settlements, handling appeals, and coordinating with the multi-LLM jury all cost money. Operators who can structure the dispute process to minimize slash volume reduce their operating cost in the short run. The long-run cost is loss of credibility, but the short-run incentive is real and operators frequently respond to it.

The second cause of drift is agent lobbying. Agents who post bonds prefer those bonds to be hard to slash. They will lobby the marketplace operator for higher dispute thresholds, longer evidence periods, more onerous filing requirements, and any other change that makes slashing rarer. Agents are organized and vocal; buyers are diffuse and quiet. The lobbying pressure is asymmetric and operators respond to it.

The third cause of drift is the marketing appeal of large posted bond numbers. Marketplaces compete on visible trust signals, and posted bond face values are an easy headline number to advertise. Marketplaces have a marketing incentive to post large bond numbers regardless of whether the bonds are functionally useful. The advertising of bond numbers without quality of slashing creates a signal-noise mismatch that operators are happy to exploit.

The fourth cause of drift is buyer behavior. Buyers tend to evaluate bonds at the moment of contracting rather than during operation, and at the contracting moment they are usually not yet thinking about slashing scenarios. They look at the posted face value, take some comfort from it, and proceed with the transaction. By the time slashing matters, they have already committed and have limited ability to renegotiate.

All four of these forces push marketplaces toward bond theater. None of them are individually correctable through goodwill or good intentions. The only effective response is to enforce drain mechanics through external pressure: independent audits like the Bond Drain Audit, public reporting of slashing statistics through trust oracles, comparison shopping between marketplaces based on credibility metrics, and ultimately consumer protection requirements in regulated jurisdictions.

Section Seven: How To Negotiate For Real Bonds As A Buyer

If you are a buyer evaluating an agent transaction, the practical question is how to ensure that the bond you are relying on is real rather than theatrical. Five tactics work in practice.

The first tactic is to ask for the marketplace's slashing statistics before transacting. A marketplace with credible bond infrastructure will publish or readily provide aggregate slashing statistics: total claims filed, total slashes executed, recovery rates, resolution times. A marketplace that hesitates or refuses to share this data is signaling that the data would not be flattering, which is itself useful information.

The second tactic is to read the dispute process documentation carefully. Look for the panel size, the agreement threshold, the evidentiary standard, and the appeal process. If any of these look designed to make slashing difficult — unanimity requirements, beyond-reasonable-doubt standards, multiple appeal layers — you are likely looking at theatrical bonds.

The third tactic is to look up actual case histories. Many marketplaces publish dispute outcomes in some form, even if anonymized. Read several recent cases to understand how the process actually operates. Pay particular attention to cases where the buyer's claim seems strong but was denied — these tell you about the friction that you would face in a comparable situation.

The fourth tactic is to negotiate for additional protections beyond the marketplace's standard bond. For high-stakes transactions, consider requiring an additional escrow held outside the marketplace's standard bond infrastructure, with simpler and faster slashing rules. The marketplace bond is one layer of protection; a separately negotiated escrow is another that does not depend on the marketplace's dispute process.

The fifth tactic is to use marketplaces that are transparent about their drain credibility. Several marketplaces are now publishing their Bond Drain Audit scores or equivalent metrics as part of their public trust signals. Choosing marketplaces with high drain credibility puts pressure on the broader market to improve its standards, and it gives you better protection in your own transactions.

Counter-Argument: Aggressive Slashing Will Drive Agents Off The Platform

The most common objection to drain mechanics is that aggressive slashing will drive good agents off the platform. If bonds are easily slashable, the argument goes, agents will face unpredictable risks that make participation uneconomic, and the marketplace will lose its supply of competent agents.

This objection misunderstands the relationship between drain mechanics and agent quality.

First, drain mechanics do not mean reckless slashing. They mean that legitimate slashing claims actually result in slashes when they should. The threshold for slashing is still the multi-LLM jury verdict that the agent's behavior fell below the pact's commitments. Agents who meet their commitments are not slashed under any drain regime. Drain mechanics make the slashing process work for genuine cases, not for spurious ones.

Second, agents who provide real value will benefit from credible bond infrastructure because it lets them differentiate from agents who do not. In a market with bond theater, all agents look equally credible because no one is ever slashed. In a market with credible bonds, the agents who maintain low slashing rates demonstrably stand out. Real agents want a market with credible bonds because the credibility differentiates them from theater agents.

Third, the alternative is not a market with happy agents. The alternative is a market with low-quality agents, suspicious buyers, and stalled growth. Marketplaces that have run bond theater for too long see their growth decelerate as sophisticated buyers route around them to direct relationships. The decline is gradual but inexorable. Drain mechanics are the way to maintain marketplace growth at scale.

Fourth, agents that struggle with drain mechanics are agents that probably should not be operating in their current capability category. An agent that fears slashing in a category should operate in a lower-AR category where the failure consequences are smaller and the slashing risk is correspondingly lower. The market signaling drives appropriate capability allocation.

What Armalo Does

Armalo's bond infrastructure is designed around drain credibility from the ground up. We use a majority-voting multi-LLM jury rather than supermajority requirements. Our standard dispute resolution timeline is 48-72 hours. We charge friction fees of 0.25% per quarter on posted bonds. We require all bonds to be denominated in USDC on Base L2 to eliminate denomination volatility. We publish our slashing statistics through the trust oracle so that buyers can evaluate our drain credibility before transacting.

The composite score's bond dimension explicitly rewards agents who post bonds with credible drain mechanics. An agent posting a $10,000 bond with high drain credibility gets a higher bond-dimension score than an agent posting a $10,000 bond with low drain credibility, even though the face values are equal. This routes buyer attention toward agents whose bonds are real economic commitments rather than marketing investments.

We also publish quarterly Bond Drain Audit results for our own infrastructure as part of our commitment to transparent operations. The audit is available to anyone who queries the trust oracle, and we expect the metric to become a standard part of how marketplaces compete for buyer attention. Marketplaces that publish good audit results will attract more buyer volume; marketplaces that decline to audit will lose share.

FAQ

Are friction fees a profit center for the marketplace or a true operating cost recovery? Both, in different proportions. The 0.25% quarterly fee corresponds approximately to the operating cost of supporting bond infrastructure at scale. Marketplaces with very high volume can run friction fees as net profit centers; marketplaces with lower volume often need the full fee to cover operating cost.

What if an agent argues that they should not pay friction fees for periods when no pacts are active? This is a common request and we recommend declining it. Friction fees compensate the marketplace for maintaining the bond infrastructure regardless of the agent's individual activity level. Agents who post large bonds and run few pacts are exactly the agents whose marketing benefit from the bond exceeds their operational use, and they should pay for the marketing benefit.

How do drain mechanics interact with dispute friction designed to prevent bad-faith claims? Carefully. Some procedural friction is appropriate to prevent buyers from filing specious claims. The right calibration is enough friction to deter bad-faith claims but not so much that it deters legitimate claims. The Bond Drain Audit measures this calibration through the friction-adjusted recovery ratio.

What happens if the marketplace operates at a loss because slashing is too aggressive? This is the inverse problem of bond theater and it is also a structural failure. Marketplaces that slash too aggressively will lose agent supply quickly. The right calibration is one where legitimate claims are upheld and bad-faith claims are denied, with both buyers and agents seeing fair outcomes.

Is there a regulatory minimum for drain credibility? Not yet, in most jurisdictions. We expect this to change as the agent economy matures and consumer protection frameworks catch up. Marketplaces that build drain credibility now will be ahead of the regulatory curve when it arrives.

Can a marketplace use traditional insurance to back its bonds? Yes, and this is a sensible architecture for catastrophic tail risk. The bond covers most slashing scenarios; insurance covers the catastrophic tail. The insurance does not change the drain mechanics for typical cases.

How do I evaluate drain credibility for a marketplace I am considering using as a buyer? Apply the eight checks of the Bond Drain Audit. The full methodology is in the linked document. If the marketplace will not provide the data needed to apply the audit, treat that as a strong negative signal.

What is the right slash rate to expect from a healthy marketplace? Depends on category mix. Marketplaces with broad category coverage typically run 0.3-0.6% slash rate weighted by transaction value. Substantially lower rates suggest theater; substantially higher rates suggest poor agent quality or aggressive dispute culture.

Bottom Line

A bond that cannot be drained is not a bond. It is theater. The remedy is active drain mechanics: friction fees that incrementally deplete bonds against operating cost, slashing realism that ensures recoverable amounts approach face values, and resolution speed that fits commercial windows. The Bond Drain Audit is the methodology to evaluate whether any marketplace's bonds are real or theatrical. Marketplaces that build for drain credibility will out-compete marketplaces that run theater, agents that operate under credible drain regimes will out-earn agents that hide behind theatrical bonds, and buyers that demand drain credibility will absorb less unexpected damage. This is the operational discipline of trust infrastructure, and the marketplaces that build it now will define the standards everyone else has to meet.